Content-Type: text/plain;
	charset=utf-8
Mime-Version: 1.0 (Mac OS X Mail 16.0 \(3864.400.21\))
Subject: Re: FOR PORTION OF does not recompute GENERATED STORED columns that
 depend on the range column
From: Chao Li <li.evan.chao@gmail.com>
In-Reply-To: 
 <CA+renyUBLAynaj0BKhajB6F=sLuitQkjT+_sOt5HBSRn82iQsw@mail.gmail.com>
Date: Fri, 8 May 2026 15:09:46 +0800
Cc: Peter Eisentraut <peter@eisentraut.org>,
 jian he <jian.universality@gmail.com>,
 SATYANARAYANA NARLAPURAM <satyanarlapuram@gmail.com>,
 PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Content-Transfer-Encoding: quoted-printable
Message-Id: <74C1863C-2C2A-423A-BDE7-0228889F1D80@gmail.com>
References: 
 <CAHg+QDcd=t69gLf9yQexO07EJ2mx0Z70NFHo6h94X1EDA=hM0g@mail.gmail.com>
 <CACJufxGreOtA-S-qeHyS5iSSsj5zZX0W3Rf8FxbyL+SVXFjLYw@mail.gmail.com>
 <CAHg+QDeGLfz8YSCChjqrxaVSrz9AnMA0NrmsNogLqeGgCt7-wg@mail.gmail.com>
 <CA+renyWD+XXifwswE74vhjooqbiVKu4qVhLvpMcUQBzrjVjT7A@mail.gmail.com>
 <CACJufxHYntqy2fo9CFWDDrqKjcMK8DGRM3kse4YnXYnPYq2Hiw@mail.gmail.com>
 <CA+renyVp4rgj8x0ERXRkZp223eyBZ_XZr2RVCXvjzKBhTtS6Yw@mail.gmail.com>
 <CACJufxEkomKYmWgqXJmQr_qS+z=BZ3w801eh7Z7ekh-3oHXxHQ@mail.gmail.com>
 <CA+renyWk7kVsZJPZKzN95mYkO7S=hDUx=+fUPtbg9qFqeepCpg@mail.gmail.com>
 <CACJufxHai+HB1gkNqVEHe4oKyUmXfAagWBYAWXYKy8hyMV3RxA@mail.gmail.com>
 <CA+renyUBQdhnYxfPay+dxFs6BU1-fnEQskT0r-3dQ2v-ZnmZzg@mail.gmail.com>
 <CA+renyVZLYSghHb_85w0pUBG0KNGKTwciFTKBK5--rpHUM+VdA@mail.gmail.com>
 <e856c439-bf96-4244-8dc3-18b2a552e406@eisentraut.org>
 <CA+renyU6rNkiNGreMyQ7pU_F6-5RND5jchHbECH4NoRO7W0Q-Q@mail.gmail.com>
 <27BD5D23-19C9-4FD1-8935-9C788C3C9869@gmail.com>
 <66C1555B-CA54-4ED1-AB4F-0EE97D24A006@gmail.com>
 <91B35E0F-5DC1-4417-A1B9-FAF4A3DCD2BD@gmail.com>
 <CA+renyUBLAynaj0BKhajB6F=sLuitQkjT+_sOt5HBSRn82iQsw@mail.gmail.com>
To: Paul A Jungwirth <pj@illuminatedcomputing.com>
Archived-At: 
 <https://www.postgresql.org/message-id/74C1863C-2C2A-423A-BDE7-0228889F1D80%40gmail.com>
Precedence: bulk


> On May 8, 2026, at 07:47, Paul A Jungwirth =
<pj@illuminatedcomputing.com> wrote:
>=20
> On Thu, May 7, 2026 at 12:06=E2=80=AFAM Chao Li =
<li.evan.chao@gmail.com> wrote:
>>> =
<v10-0001-Fix-some-problems-with-UPDATE-FOR-PORTION-OF.patch><v10-0002-Fix=
-FOR-PORTION-OF-on-inherited-children-with-di.patch>
>>=20
>> A few comments for 0001:
>>=20
>> 1 - execUtils.c
>> The comment explicitly says that it is unsafe to mutate perminfo, but =
bms_add_member() does not always allocate a new bitmapset. So if =
updatedCols still points to perminfo->updatedCols, then bms_add_member() =
may mutate perminfo->updatedCols.
>>=20
>> To verify that, I added Assert(updatedCols !=3D =
perminfo->updatedCols); right after the bms_add_member(), then ran "make =
check". A lot of tests failed, which seems to confirm that =
perminfo->updatedCols was being mutated.
>>=20
>> So, I think we should make a copy the bitmapset before bms_add_member =
when needed, to make sure perminfo is not mutated, something like:
>> ```
>>                        if (updatedCols =3D=3D perminfo->updatedCols)
>>                                updatedCols =3D bms_copy(updatedCols);
>>=20
>>                        updatedCols =3D
>>                                bms_add_member(updatedCols,
>>                                                           rangeAttno =
- FirstLowInvalidHeapAttributeNumber);
>> ```
>=20
> Ah, thanks for catching this! Fixed.
>=20
>> 2 - execUtils.c
>> ```
>> +  * because the user does not need UPDATE permission on it. Now =
manualy
>> ```
>>=20
>> Typo: manualy -> manually
>=20
> Fixed.
>=20
>> 3 - nodeModifyTable.c
>> ```
>> +               /*
>> +                * If we don't have a ForPortionOfState yet, we must =
be a partition
>> +                * child being hit for the first time. Make a copy =
from the root, with
>> +                * our own TupleTableSlot. We do this lazily so that =
we don't pay the
>> +                * price of unused partitions.
>> +                */
>> +               if ((((ModifyTable *) =
context.mtstate->ps.plan)->forPortionOf) &&
>> +                       !resultRelInfo->ri_forPortionOf)
>> +               {
>> +                       ExecInitForPortionOf(context.mtstate, estate, =
resultRelInfo);
>> +               }
>> ```
>>=20
>> I think this comment is stale. It could be a partition child or an =
inheritance child.
>=20
> Okay.
>=20
>> 4 - nodeModifyTable.c
>> ```
>> +       /* Each partition needs a slot matching its tuple descriptor =
*/
>> +       leafState->fp_Existing =3D
>> +               table_slot_create(resultRelInfo->ri_RelationDesc,
>> +                                                 =
&mtstate->ps.state->es_tupleTable);
>> ```
>>=20
>> I think the comment should say "each child relation" rather than =
"each partition".
>=20
> Okay.
>=20
> In these v11 patches I've tried to separate (1) the fix for GENERATED
> STORED columns and UPDATE OF triggers (2) fixing inheritance and (and
> partitions too, for the bugs in #1). I understand why jian he combined
> these into one patch: there is some overlap. If you don't like my
> separation, let me know.
>=20
> Yours,
>=20
> --=20
> Paul              ~{:-)
> pj@illuminatedcomputing.com
> =
<v11-0001-Fix-FOR-PORTION-OF-column-dependency-tracking.patch><v11-0002-Fi=
x-FOR-PORTION-OF-with-partitions-and-inheritan.patch>

Thanks for updating the patch and making the separation. After reading =
v11, I still have a few comments for 0001.

```
+	if (relinfo->ri_forPortionOf)
+	{
+		AttrNumber  rangeAttno =3D =
relinfo->ri_forPortionOf->fp_rangeAttno;
+
+		if (!bms_is_member(rangeAttno - =
FirstLowInvalidHeapAttributeNumber,
+						   updatedCols))
+		{
+			MemoryContext oldContext;
+
+			oldContext =3D =
MemoryContextSwitchTo(estate->es_query_cxt);
+
+			updatedCols =3D bms_copy(updatedCols);
+			updatedCols =3D
+				bms_add_member(updatedCols,
+							   rangeAttno - =
FirstLowInvalidHeapAttributeNumber);
+
+			MemoryContextSwitchTo(oldContext);
+		}
 	}
```

1. I don=E2=80=99t think we should unconditionally do bms_copy, only if =
(updatedCols =3D=3D perminfo->updatedCols), we need to make the copy.

2. I doubt if we need to switch to estate->es_query_cxt. Because =
ExecGetUpdatedCols() is called by ExecGetAllUpdatedCols(), and its =
header comment says the function runs in per-tuple memory context:
```
/*
* Return columns being updated, including generated columns
*
* The bitmap is allocated in per-tuple memory context. It's up to the =
caller to
* copy it into a different context with the appropriate lifespan, if =
needed.
*/
Bitmapset *
ExecGetAllUpdatedCols(ResultRelInfo *relinfo, EState *estate)
```

So I think bms_copy and bms_add_member should be just done in the =
current memory context.

3. "rangeAttno - FirstLowInvalidHeapAttributeNumber=E2=80=9D appears =
twice, maybe add a local variable to avoid the duplication.

Best regards,
--
Chao Li (Evan)
HighGo Software Co., Ltd.
https://www.highgo.com/