agora inbox for [email protected]  

help / color / mirror / Atom feed

[PATCH v2 3/3] remove support for PQfn
334+ messages / 2 participants
[nested] [flat]

* [PATCH v2 3/3] remove support for PQfn
@ 2026-05-22 19:00  Nathan Bossart <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Nathan Bossart @ 2026-05-22 19:00 UTC (permalink / raw)

---
 doc/src/sgml/libpq.sgml             | 119 +-------------
 src/backend/tcop/fastpath.c         |   3 +-
 src/include/tcop/dest.h             |   4 +-
 src/interfaces/libpq/fe-exec.c      |  54 +-----
 src/interfaces/libpq/fe-protocol3.c | 246 ----------------------------
 src/interfaces/libpq/libpq-int.h    |   5 -
 6 files changed, 12 insertions(+), 419 deletions(-)

diff --git a/doc/src/sgml/libpq.sgml b/doc/src/sgml/libpq.sgml
index 7d3c3bb66d8..812e9089bfd 100644
--- a/doc/src/sgml/libpq.sgml
+++ b/doc/src/sgml/libpq.sgml
@@ -3738,7 +3738,7 @@ PGresult *PQdescribePrepared(PGconn *conn, const char *stmtName);
         <xref linkend="libpq-PQparamtype"/> can be applied to this
         <structname>PGresult</structname> to obtain information about the parameters
         of the prepared statement, and the functions
-        <xref linkend="libpq-PQnfields"/>, <xref linkend="libpq-PQfname"/>,
+        <xref linkend="libpq-PQnfields"/>,
         <xref linkend="libpq-PQftype"/>, etc. provide information about the
         result columns (if any) of the statement.
        </para>
@@ -5887,7 +5887,7 @@ int PQflush(PGconn *conn);
     are permitted, command strings containing multiple SQL commands are
     disallowed, and so is <literal>COPY</literal>.
     Using synchronous command execution functions
-    such as <function>PQfn</function>,
+    such as
     <function>PQexec</function>,
     <function>PQexecParams</function>,
     <function>PQprepare</function>,
@@ -7046,121 +7046,6 @@ int PQrequestCancel(PGconn *conn);
   </sect2>
  </sect1>
 
- <sect1 id="libpq-fastpath">
-  <title>The Fast-Path Interface</title>
-
-  <indexterm zone="libpq-fastpath">
-   <primary>fast path</primary>
-  </indexterm>
-
-  <para>
-   <productname>PostgreSQL</productname> provides a fast-path interface
-   to send simple function calls to the server.
-  </para>
-
-  <warning>
-   <para>
-    This interface is unsafe and should not be used.  When
-    <parameter>result_is_int</parameter> is set to <literal>0</literal>,
-    <function>PQfn</function> may write data beyond the end of
-    <parameter>result_buf</parameter>, regardless of whether the buffer has
-    enough space for the requested number of bytes.  Furthermore, it is
-    obsolete, as one can achieve similar
-    performance and greater functionality by setting up a prepared
-    statement to define the function call.  Then, executing the statement
-    with binary transmission of parameters and results substitutes for a
-    fast-path function call.
-   </para>
-  </warning>
-
-  <para>
-   The function <function id="libpq-PQfn">PQfn</function><indexterm><primary>PQfn</primary></indexterm>
-   requests execution of a server function via the fast-path interface:
-<synopsis>
-PGresult *PQfn(PGconn *conn,
-               int fnid,
-               int *result_buf,
-               int *result_len,
-               int result_is_int,
-               const PQArgBlock *args,
-               int nargs);
-
-typedef struct
-{
-    int len;
-    int isint;
-    union
-    {
-        int *ptr;
-        int integer;
-    } u;
-} PQArgBlock;
-</synopsis>
-  </para>
-
-  <para>
-   The <parameter>fnid</parameter> argument is the OID of the function to be
-   executed.  <parameter>args</parameter> and <parameter>nargs</parameter> define the
-   parameters to be passed to the function; they must match the declared
-   function argument list.  When the <parameter>isint</parameter> field of a
-   parameter structure is true, the <parameter>u.integer</parameter> value is sent
-   to the server as an integer of the indicated length (this must be
-   2 or 4 bytes); proper byte-swapping occurs.  When <parameter>isint</parameter>
-   is false, the indicated number of bytes at <parameter>*u.ptr</parameter> are
-   sent with no processing; the data must be in the format expected by
-   the server for binary transmission of the function's argument data
-   type.  (The declaration of <parameter>u.ptr</parameter> as being of
-   type <type>int *</type> is historical; it would be better to consider
-   it <type>void *</type>.)
-   <parameter>result_buf</parameter> points to the buffer in which to place
-   the function's return value.  The caller must have allocated sufficient
-   space to store the return value.  (There is no check!) The actual result
-   length in bytes will be returned in the integer pointed to by
-   <parameter>result_len</parameter>.  If a 2- or 4-byte integer result
-   is expected, set <parameter>result_is_int</parameter> to 1, otherwise
-   set it to 0.  Setting <parameter>result_is_int</parameter> to 1 causes
-   <application>libpq</application> to byte-swap the value if necessary, so that it
-   is delivered as a proper <type>int</type> value for the client machine;
-   note that a 4-byte integer is delivered into <parameter>*result_buf</parameter>
-   for either allowed result size.
-   When <parameter>result_is_int</parameter> is 0, the binary-format byte string
-   sent by the server is returned unmodified. (In this case it's better
-   to consider <parameter>result_buf</parameter> as being of
-   type <type>void *</type>.)
-  </para>
-
-  <para>
-   <function>PQfn</function> always returns a valid
-   <structname>PGresult</structname> pointer, with
-   status <literal>PGRES_COMMAND_OK</literal> for success
-   or <literal>PGRES_FATAL_ERROR</literal> if some problem was encountered.
-   The result status should be
-   checked before the result is used.   The caller is responsible for
-   freeing  the  <structname>PGresult</structname>  with
-   <xref linkend="libpq-PQclear"/> when it is no longer needed.
-  </para>
-
-  <para>
-   To pass a NULL argument to the function, set
-   the <parameter>len</parameter> field of that parameter structure
-   to <literal>-1</literal>; the <parameter>isint</parameter>
-   and <parameter>u</parameter> fields are then irrelevant.
-  </para>
-
-  <para>
-   If the function returns NULL, <parameter>*result_len</parameter> is set
-   to <literal>-1</literal>, and <parameter>*result_buf</parameter> is not
-   modified.
-  </para>
-
-  <para>
-   Note that it is not possible to handle set-valued results when using
-   this interface.  Also, the function must be a plain function, not an
-   aggregate, window function, or procedure.
-  </para>
-
- </sect1>
-
  <sect1 id="libpq-notify">
   <title>Asynchronous Notification</title>
 
diff --git a/src/backend/tcop/fastpath.c b/src/backend/tcop/fastpath.c
index 52772bc90a8..2a5c45efffe 100644
--- a/src/backend/tcop/fastpath.c
+++ b/src/backend/tcop/fastpath.c
@@ -11,7 +11,8 @@
  *	  src/backend/tcop/fastpath.c
  *
  * NOTES
- *	  This cruft is the server side of PQfn.
+ *	  This cruft is the server side of PQfn, which was removed in v20 but
+ *	  may still be used by older clients.
  *
  *-------------------------------------------------------------------------
  */
diff --git a/src/include/tcop/dest.h b/src/include/tcop/dest.h
index 103f27fc3cb..507414421ec 100644
--- a/src/include/tcop/dest.h
+++ b/src/include/tcop/dest.h
@@ -12,8 +12,8 @@
  *
  *	  - a remote process is the destination when we are
  *		running a backend with a frontend and the frontend executes
- *		PQexec() or PQfn().  In this case, the results are sent
- *		to the frontend via the functions in backend/libpq.
+ *		PQexec().  In this case, the results are sent to the frontend via
+ *		the functions in backend/libpq.
  *
  *	  - DestNone is the destination when the system executes
  *		a query internally.  The results are discarded.
diff --git a/src/interfaces/libpq/fe-exec.c b/src/interfaces/libpq/fe-exec.c
index 7b8edacbfde..400e1eef94b 100644
--- a/src/interfaces/libpq/fe-exec.c
+++ b/src/interfaces/libpq/fe-exec.c
@@ -2986,13 +2986,11 @@ PQendcopy(PGconn *conn)
  *		nargs			: # of arguments in args array.
  *
  * RETURNS
- *		PGresult with status = PGRES_COMMAND_OK if successful.
- *			*result_len is > 0 if there is a return value, 0 if not.
- *		PGresult with status = PGRES_FATAL_ERROR if backend returns an error.
- *		NULL on communications failure.  conn->errorMessage will be set.
+ *		This function was unsafe and is no longer supported, so it now always
+ * 		sets *result_len to 0 and returns a PGresult with status set to
+ *		PGRES_FATAL_ERROR.
  * ----------------
  */
-
 PGresult *
 PQfn(PGconn *conn,
 	 int fnid,
@@ -3001,51 +2999,11 @@ PQfn(PGconn *conn,
 	 int result_is_int,
 	 const PQArgBlock *args,
 	 int nargs)
-{
-	return PQnfn(conn, fnid, result_buf, -1, result_len,
-				 result_is_int, args, nargs);
-}
-
-/*
- * PQnfn
- *		Private version of PQfn() with verification that returned data fits in
- *		result_buf when result_is_int == 0.  Setting buf_size to -1 disables
- *		this verification.
- */
-PGresult *
-PQnfn(PGconn *conn, int fnid, int *result_buf, int buf_size, int *result_len,
-	  int result_is_int, const PQArgBlock *args, int nargs)
 {
 	*result_len = 0;
-
-	if (!conn)
-		return NULL;
-
-	/*
-	 * Since this is the beginning of a query cycle, reset the error state.
-	 * However, in pipeline mode with something already queued, the error
-	 * buffer belongs to that command and we shouldn't clear it.
-	 */
-	if (conn->cmd_queue_head == NULL)
-		pqClearConnErrorState(conn);
-
-	if (conn->pipelineStatus != PQ_PIPELINE_OFF)
-	{
-		libpq_append_conn_error(conn, "%s not allowed in pipeline mode", "PQfn");
-		return NULL;
-	}
-
-	if (conn->sock == PGINVALID_SOCKET || conn->asyncStatus != PGASYNC_IDLE ||
-		pgHavePendingResult(conn))
-	{
-		libpq_append_conn_error(conn, "connection in wrong state");
-		return NULL;
-	}
-
-	return pqFunctionCall3(conn, fnid,
-						   result_buf, buf_size, result_len,
-						   result_is_int,
-						   args, nargs);
+	libpq_append_conn_error(conn, "PQfn() is no longer supported");
+	pqSaveErrorResult(conn);
+	return pqPrepareAsyncResult(conn);
 }
 
 /* ====== Pipeline mode support ======== */
diff --git a/src/interfaces/libpq/fe-protocol3.c b/src/interfaces/libpq/fe-protocol3.c
index 0407d10362d..f9c55ee7905 100644
--- a/src/interfaces/libpq/fe-protocol3.c
+++ b/src/interfaces/libpq/fe-protocol3.c
@@ -2235,252 +2235,6 @@ pqEndcopy3(PGconn *conn)
 	return 1;
 }
 
-
-/*
- * PQfn - Send a function call to the POSTGRES backend.
- *
- * See fe-exec.c for documentation.
- */
-PGresult *
-pqFunctionCall3(PGconn *conn, Oid fnid,
-				int *result_buf, int buf_size, int *actual_result_len,
-				int result_is_int,
-				const PQArgBlock *args, int nargs)
-{
-	bool		needInput = false;
-	ExecStatusType status = PGRES_FATAL_ERROR;
-	char		id;
-	int			msgLength;
-	int			avail;
-	int			i;
-
-	/* already validated by PQfn */
-	Assert(conn->pipelineStatus == PQ_PIPELINE_OFF);
-
-	/* PQfn already validated connection state */
-
-	if (pqPutMsgStart(PqMsg_FunctionCall, conn) < 0 ||
-		pqPutInt(fnid, 4, conn) < 0 ||	/* function id */
-		pqPutInt(1, 2, conn) < 0 || /* # of format codes */
-		pqPutInt(1, 2, conn) < 0 || /* format code: BINARY */
-		pqPutInt(nargs, 2, conn) < 0)	/* # of args */
-	{
-		/* error message should be set up already */
-		return NULL;
-	}
-
-	for (i = 0; i < nargs; ++i)
-	{							/* len.int4 + contents	   */
-		if (pqPutInt(args[i].len, 4, conn))
-			return NULL;
-		if (args[i].len == -1)
-			continue;			/* it's NULL */
-
-		if (args[i].isint)
-		{
-			if (pqPutInt(args[i].u.integer, args[i].len, conn))
-				return NULL;
-		}
-		else
-		{
-			if (pqPutnchar(args[i].u.ptr, args[i].len, conn))
-				return NULL;
-		}
-	}
-
-	if (pqPutInt(1, 2, conn) < 0)	/* result format code: BINARY */
-		return NULL;
-
-	if (pqPutMsgEnd(conn) < 0 ||
-		pqFlush(conn))
-		return NULL;
-
-	for (;;)
-	{
-		if (needInput)
-		{
-			/* Wait for some data to arrive (or for the channel to close) */
-			if (pqWait(true, false, conn) ||
-				pqReadData(conn) < 0)
-				break;
-		}
-
-		/*
-		 * Scan the message. If we run out of data, loop around to try again.
-		 */
-		needInput = true;
-
-		conn->inCursor = conn->inStart;
-		if (pqGetc(&id, conn))
-			continue;
-		if (pqGetInt(&msgLength, 4, conn))
-			continue;
-
-		/*
-		 * Try to validate message type/length here.  A length less than 4 is
-		 * definitely broken.  Large lengths should only be believed for a few
-		 * message types.
-		 */
-		if (msgLength < 4)
-		{
-			handleSyncLoss(conn, id, msgLength);
-			break;
-		}
-		if (msgLength > 30000 && !VALID_LONG_MESSAGE_TYPE(id))
-		{
-			handleSyncLoss(conn, id, msgLength);
-			break;
-		}
-
-		/*
-		 * Can't process if message body isn't all here yet.
-		 */
-		msgLength -= 4;
-		avail = conn->inEnd - conn->inCursor;
-		if (avail < msgLength)
-		{
-			/*
-			 * Before looping, enlarge the input buffer if needed to hold the
-			 * whole message.  See notes in parseInput.
-			 */
-			if (pqCheckInBufferSpace(conn->inCursor + (size_t) msgLength,
-									 conn))
-			{
-				/*
-				 * Abandon the connection.  There's not much else we can
-				 * safely do; we can't just ignore the message or we could
-				 * miss important changes to the connection state.
-				 * pqCheckInBufferSpace() already reported the error.
-				 */
-				handleFatalError(conn);
-				break;
-			}
-			continue;
-		}
-
-		/*
-		 * We should see V or E response to the command, but might get N
-		 * and/or A notices first. We also need to swallow the final Z before
-		 * returning.
-		 */
-		switch (id)
-		{
-			case PqMsg_FunctionCallResponse:
-				if (pqGetInt(actual_result_len, 4, conn))
-					continue;
-				if (*actual_result_len != -1)
-				{
-					if (result_is_int)
-					{
-						if (pqGetInt(result_buf, *actual_result_len, conn))
-							continue;
-					}
-					else
-					{
-						/*
-						 * If the server returned too much data for the
-						 * buffer, something fishy is going on.  Abandon ship.
-						 */
-						if (buf_size != -1 && *actual_result_len > buf_size)
-						{
-							libpq_append_conn_error(conn, "server returned too much data");
-							handleFatalError(conn);
-							return pqPrepareAsyncResult(conn);
-						}
-
-						if (pqGetnchar(result_buf,
-									   *actual_result_len,
-									   conn))
-							continue;
-					}
-				}
-				/* correctly finished function result message */
-				status = PGRES_COMMAND_OK;
-				break;
-			case PqMsg_ErrorResponse:
-				if (pqGetErrorNotice3(conn, true))
-					continue;
-				status = PGRES_FATAL_ERROR;
-				break;
-			case PqMsg_NotificationResponse:
-				/* handle notify and go back to processing return values */
-				if (getNotify(conn))
-					continue;
-				break;
-			case PqMsg_NoticeResponse:
-				/* handle notice and go back to processing return values */
-				if (pqGetErrorNotice3(conn, false))
-					continue;
-				break;
-			case PqMsg_ReadyForQuery:
-				if (getReadyForQuery(conn))
-					continue;
-
-				/* consume the message */
-				pqParseDone(conn, conn->inStart + 5 + msgLength);
-
-				/*
-				 * If we already have a result object (probably an error), use
-				 * that.  Otherwise, if we saw a function result message,
-				 * report COMMAND_OK.  Otherwise, the backend violated the
-				 * protocol, so complain.
-				 */
-				if (!pgHavePendingResult(conn))
-				{
-					if (status == PGRES_COMMAND_OK)
-					{
-						conn->result = PQmakeEmptyPGresult(conn, status);
-						if (!conn->result)
-						{
-							libpq_append_conn_error(conn, "out of memory");
-							pqSaveErrorResult(conn);
-						}
-					}
-					else
-					{
-						libpq_append_conn_error(conn, "protocol error: no function result");
-						pqSaveErrorResult(conn);
-					}
-				}
-				/* and we're out */
-				return pqPrepareAsyncResult(conn);
-			case PqMsg_ParameterStatus:
-				if (getParameterStatus(conn))
-					continue;
-				break;
-			case PqMsg_PrepStmtDeallocated:
-				if (getPrepStmtDeallocated(conn))
-					continue;
-				break;
-			default:
-				/* The backend violates the protocol. */
-				libpq_append_conn_error(conn, "protocol error: id=0x%x", id);
-				pqSaveErrorResult(conn);
-
-				/*
-				 * We can't call parsing done due to the protocol violation
-				 * (so message tracing wouldn't work), but trust the specified
-				 * message length as what to skip.
-				 */
-				conn->inStart += 5 + msgLength;
-				return pqPrepareAsyncResult(conn);
-		}
-
-		/* Completed parsing this message, keep going */
-		pqParseDone(conn, conn->inStart + 5 + msgLength);
-		needInput = false;
-	}
-
-	/*
-	 * We fall out of the loop only upon failing to read data.
-	 * conn->errorMessage has been set by pqWait or pqReadData. We want to
-	 * append it to any already-received error message.
-	 */
-	pqSaveErrorResult(conn);
-	return pqPrepareAsyncResult(conn);
-}
-
-
 /*
  * Construct startup packet
  *
diff --git a/src/interfaces/libpq/libpq-int.h b/src/interfaces/libpq/libpq-int.h
index 7a32c14de5e..980ef5374b5 100644
--- a/src/interfaces/libpq/libpq-int.h
+++ b/src/interfaces/libpq/libpq-int.h
@@ -768,11 +768,6 @@ extern int	pqGetCopyData3(PGconn *conn, char **buffer, int async);
 extern int	pqGetline3(PGconn *conn, char *s, int maxlen);
 extern int	pqGetlineAsync3(PGconn *conn, char *buffer, int bufsize);
 extern int	pqEndcopy3(PGconn *conn);
-extern PGresult *pqFunctionCall3(PGconn *conn, Oid fnid,
-								 int *result_buf, int buf_size,
-								 int *actual_result_len,
-								 int result_is_int,
-								 const PQArgBlock *args, int nargs);
 
 /* === in fe-cancel.c === */
 
-- 
2.50.1 (Apple Git-155)


--wfzLmV/Wl0ehOJBp--





^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread

* [PATCH 1/8] Use tuple slot to pass tuples for rewriting.
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 334+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

This patch tries to adopt the preferable way of handling tuples, i.e. pass the
containing tuple slot rather than the actual tuple. The motivation is that
heap_insert_for_repack() will need a slot in the near future, in order to call
ExecInsertIndexTuples(). Also, in order to set values of dropped attributes to
to NULL, it seems more compact to pass a tuple slot as a workspace than two
arrays (one for values and one for nulls).
---
 src/backend/access/heap/heapam_handler.c | 199 +++++++++++++----------
 1 file changed, 110 insertions(+), 89 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 2268cc277bc..127d4415084 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -48,14 +48,13 @@
 #include "utils/rel.h"
 #include "utils/tuplesort.h"
 
-static void reform_and_rewrite_tuple(HeapTuple tuple,
-									 Relation OldHeap, Relation NewHeap,
-									 Datum *values, bool *isnull, RewriteState rwstate);
-static void heap_insert_for_repack(HeapTuple tuple, Relation OldHeap,
-								   Relation NewHeap, Datum *values, bool *isnull,
-								   BulkInsertState bistate);
-static HeapTuple reform_tuple(HeapTuple tuple, Relation OldHeap,
-							  Relation NewHeap, Datum *values, bool *isnull);
+static void reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+									 RewriteState rwstate);
+static void heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+								   TupleTableSlot *reform,
+								   BulkInsertStateData *bistate);
+static bool tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc);
+static void clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform);
 
 static bool SampleHeapTupleVisible(TableScanDesc scan, Buffer buffer,
 								   HeapTuple tuple,
@@ -609,11 +608,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	bool		is_system_catalog;
 	Tuplesortstate *tuplesort;
 	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
 	TupleTableSlot *slot;
-	int			natts;
-	Datum	   *values;
-	bool	   *isnull;
+	TupleTableSlot *reform_slot;
 	BufferHeapTupleTableSlot *hslot;
 	BlockNumber prev_cblock = InvalidBlockNumber;
 	bool		concurrent = snapshot != NULL;
@@ -627,11 +623,6 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 	 */
 	Assert(RelationGetTargetBlock(NewHeap) == InvalidBlockNumber);
 
-	/* Preallocate values/isnull arrays */
-	natts = newTupDesc->natts;
-	values = palloc_array(Datum, natts);
-	isnull = palloc_array(bool, natts);
-
 	/*
 	 * In non-concurrent mode, initialize the rewrite operation.  This is not
 	 * needed in concurrent mode.
@@ -705,6 +696,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 
 	slot = table_slot_create(OldHeap, NULL);
 	hslot = (BufferHeapTupleTableSlot *) slot;
+	reform_slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										   &TTSOpsVirtual);
 
 	/*
 	 * Scan through the OldHeap, either in OldIndex order or sequentially;
@@ -881,11 +874,9 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			int64		ct_val[2];
 
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple, OldHeap, NewHeap,
-										 values, isnull, rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/*
 			 * In indexscan mode and also VACUUM FULL, report increase in
@@ -901,8 +892,7 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		index_endscan(indexScan);
 	if (tableScan != NULL)
 		table_endscan(tableScan);
-	if (slot)
-		ExecDropSingleTupleTableSlot(slot);
+	ExecDropSingleTupleTableSlot(slot);
 
 	/*
 	 * In scan-and-sort mode, complete the sort, then read out all live tuples
@@ -922,6 +912,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 		pgstat_progress_update_param(PROGRESS_REPACK_PHASE,
 									 PROGRESS_REPACK_PHASE_WRITE_NEW_HEAP);
 
+		slot = MakeSingleTupleTableSlot(RelationGetDescr(OldHeap),
+										&TTSOpsHeapTuple);
 		for (;;)
 		{
 			HeapTuple	tuple;
@@ -932,33 +924,36 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
 			if (tuple == NULL)
 				break;
 
+			/*
+			 * XXX Ideally we should use tuplesort_gettupleslot() above, but
+			 * it retrieves minimal tuples and tuplesort_puttupleslot() cannot
+			 * get the tuple descriptor from tuplesort created by
+			 * tuplesort_begin_cluster().
+			 */
+			ExecStoreHeapTuple(tuple, slot, false);
+
 			n_tuples += 1;
 			if (!concurrent)
-				reform_and_rewrite_tuple(tuple,
-										 OldHeap, NewHeap,
-										 values, isnull,
-										 rwstate);
+				reform_and_rewrite_tuple(slot, reform_slot, rwstate);
 			else
-				heap_insert_for_repack(tuple, OldHeap, NewHeap,
-									   values, isnull, bistate);
+				heap_insert_for_repack(NewHeap, slot, reform_slot, bistate);
 
 			/* Report n_tuples */
 			pgstat_progress_update_param(PROGRESS_REPACK_HEAP_TUPLES_INSERTED,
 										 n_tuples);
 		}
 
+		ExecDropSingleTupleTableSlot(slot);
 		tuplesort_end(tuplesort);
 	}
 
+	ExecDropSingleTupleTableSlot(reform_slot);
+
 	/* Write out any remaining tuples, and fsync if needed */
 	if (rwstate)
 		end_heap_rewrite(rwstate);
 	if (bistate)
 		FreeBulkInsertState(bistate);
-
-	/* Clean up */
-	pfree(values);
-	pfree(isnull);
 }
 
 /*
@@ -2346,21 +2341,45 @@ heapam_scan_sample_next_tuple(TableScanDesc scan, SampleScanState *scanstate,
  * currently only known to happen as an after-effect of ALTER TABLE
  * SET WITHOUT OIDS.
  *
- * So, we must reconstruct the tuple from component Datums.
+ * So, we must reconstruct the tuple from component Datums. 'reform' slot
+ * is a workspace for this reconstruction.
  */
 static void
-reform_and_rewrite_tuple(HeapTuple tuple,
-						 Relation OldHeap, Relation NewHeap,
-						 Datum *values, bool *isnull, RewriteState rwstate)
+reform_and_rewrite_tuple(TupleTableSlot *src, TupleTableSlot *reform,
+						 RewriteState rwstate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple,
+				newtuple;
+	bool		shouldFree,
+				shouldFreeNew;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	/*
+	 * The old tuple will not be modified, so do not request materialization.
+	 * (A copy can be created for specific slot type though, e.g.
+	 * TTSOpsMinimalTuple.)
+	 */
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+
+		/* No need to materialize, copy will be created anyway. */
+		Assert(TTS_IS_VIRTUAL(reform));
+		newtuple = ExecFetchSlotHeapTuple(reform, false, &shouldFreeNew);
+	}
+	else
+	{
+		newtuple = heap_copytuple(tuple);
+		shouldFreeNew = true;
+	}
 
 	/* The heap rewrite module does the rest */
 	rewrite_heap_tuple(rwstate, tuple, newtuple);
 
-	heap_freetuple(newtuple);
+	if (shouldFree)
+		heap_freetuple(tuple);
+	if (shouldFreeNew)
+		heap_freetuple(newtuple);
 }
 
 /*
@@ -2372,6 +2391,9 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * information). Thus we must use heap_insert() both during the
  * catch-up and here.
  *
+ * 'reform' is a slot to use for tuple "reforming", typically to get set
+ * values of dropped columns to NULL.
+ *
  * We pass the NO_LOGICAL flag to heap_insert() in order to skip logical
  * decoding: as soon as REPACK CONCURRENTLY swaps the relation files, it drops
  * this relation, so no logical replication subscription should need the data.
@@ -2380,76 +2402,75 @@ reform_and_rewrite_tuple(HeapTuple tuple,
  * case.
  */
 static void
-heap_insert_for_repack(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-					   Datum *values, bool *isnull, BulkInsertState bistate)
+heap_insert_for_repack(Relation rel, TupleTableSlot *src,
+					   TupleTableSlot *reform, BulkInsertStateData *bistate)
 {
-	HeapTuple	newtuple;
+	HeapTuple	tuple;
+	bool		shouldFree;
+	TupleTableSlot *slot;
 
-	newtuple = reform_tuple(tuple, OldHeap, NewHeap, values, isnull);
+	tuple = ExecFetchSlotHeapTuple(src, false, &shouldFree);
+	if (tuple_needs_reform(tuple, src->tts_tupleDescriptor))
+	{
+		clear_dropped_attributes(tuple, reform);
+		slot = reform;
+	}
+	else
+		slot = src;
 
-	heap_insert(NewHeap, newtuple, GetCurrentCommandId(true),
-				HEAP_INSERT_NO_LOGICAL, bistate);
+	/*
+	 * clear_dropped_attributes() should have deformed the tuple, so nothing
+	 * should depend on it now.
+	 */
+	if (shouldFree)
+		heap_freetuple(tuple);
 
-	heap_freetuple(newtuple);
+	table_tuple_insert(rel, slot, GetCurrentCommandId(true),
+					   TABLE_INSERT_NO_LOGICAL, bistate);
 }
 
-/*
- * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
- *
- * Deform the given tuple, set values of dropped columns to NULL, and fill in
- * any values from attmissingval; then form a new tuple and return it.  If no
- * attributes need to be changed, a copy of the original tuple is returned.
- * Caller is responsible for freeing the returned tuple.
- *
- * XXX this coding assumes that both relations have the same tupledesc.
- */
-static HeapTuple
-reform_tuple(HeapTuple tuple, Relation OldHeap, Relation NewHeap,
-			 Datum *values, bool *isnull)
+static bool
+tuple_needs_reform(HeapTuple tuple, TupleDesc tupDesc)
 {
-	TupleDesc	oldTupDesc = RelationGetDescr(OldHeap);
-	TupleDesc	newTupDesc = RelationGetDescr(NewHeap);
-	bool		needs_reform = false;
-
 	/*
 	 * A short tuple might require values from attmissing val, so activate the
 	 * coding unconditionally in that case.  The value might legitimally be
 	 * NULL otherwise, so this is slightly wasteful, but it probably beats
 	 * having to test each attribute for presence of attmissingval each time.
 	 */
-	if (HeapTupleHeaderGetNatts(tuple->t_data) < newTupDesc->natts)
-		needs_reform = true;
+	if (HeapTupleHeaderGetNatts(tuple->t_data) < tupDesc->natts)
+		return true;
 
-	/*
-	 * If the column has been dropped but a value is still present, we can
-	 * optimize storage now by getting rid of it.
-	 */
-	if (!needs_reform)
+	/* Does it have dropped attributes? */
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		for (int i = 0; i < newTupDesc->natts; i++)
-		{
-			if (TupleDescCompactAttr(newTupDesc, i)->attisdropped &&
-				!heap_attisnull(tuple, i + 1, newTupDesc))
-			{
-				needs_reform = true;
-				break;
-			}
-		}
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped &&
+			!heap_attisnull(tuple, i + 1, tupDesc))
+			return true;
 	}
 
-	/* Skip work if no changes are needed */
-	if (!needs_reform)
-		return heap_copytuple(tuple);
+	return false;
+}
 
-	heap_deform_tuple(tuple, oldTupDesc, values, isnull);
+/*
+ * Subroutine for reform_and_rewrite_tuple and heap_insert_for_repack.
+ *
+ * Set values of dropped columns to NULL,
+ */
+static void
+clear_dropped_attributes(HeapTuple tuple, TupleTableSlot *reform)
+{
+	TupleDesc	tupDesc = reform->tts_tupleDescriptor;
+
+	/* Assuming 'reform' is virtual, this deforms the tuple. */
+	Assert(TTS_IS_VIRTUAL(reform));
+	ExecForceStoreHeapTuple(tuple, reform, false);
 
-	for (int i = 0; i < newTupDesc->natts; i++)
+	for (int i = 0; i < tupDesc->natts; i++)
 	{
-		if (TupleDescCompactAttr(newTupDesc, i)->attisdropped)
-			isnull[i] = true;
+		if (TupleDescCompactAttr(tupDesc, i)->attisdropped)
+			reform->tts_isnull[i] = true;
 	}
-
-	return heap_form_tuple(newTupDesc, values, isnull);
 }
 
 /*
-- 
2.52.0


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=v01-0002-Move-functions-to-repack.c.patch



^ permalink  raw  reply  [nested|flat] 334+ messages in thread