Message-ID: <51985D31.1030705@2ndQuadrant.com>
Date: Sun, 19 May 2013 01:03:45 -0400
From: Greg Smith <greg@2ndQuadrant.com>
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.7;
	rv:14.0) Gecko/20120713 Thunderbird/14.0
MIME-Version: 1.0
To: pgsql-docs <pgsql-docs@postgresql.org>
Subject: Clarify TIMING OFF behavior
Content-Type: multipart/mixed; boundary="------------030400060101080808050004"
Precedence: bulk
Sender: pgsql-docs-owner@postgresql.org

This is a multi-part message in MIME format.
--------------030400060101080808050004
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

While chatting in #pedant today, I noticed that it's not clear that 
EXPLAIN (ANALYZE ON, TIMING OFF) will still collect the total runtime of 
the statement.  Attached patch clarifies that.  It also comments why the 
related code runs INSTR_TIME_SET_CURRENT in all cases.  That's actually 
how I noticed that--a read through explain.c now could rightly wonder 
"why is it still collecting that bit of time info when TIMING is off?"

There's an argument possible here that a third TIMING option besides 
ON/OFF should really exist, one that really doesn't do any timing at 
all.  I don't think enough use cases for that exist to justify coding 
it.  The low overhead combination I've wanted for a while is EXPLAIN 
(ANALYZE ON, BUFFERS ON, TIMING OFF), and that works.

-- 
Greg Smith   2ndQuadrant US    greg@2ndQuadrant.com   Baltimore, MD
PostgreSQL Training, Services, and 24x7 Support www.2ndQuadrant.com

--------------030400060101080808050004
Content-Type: text/plain; charset=UTF-8; x-mac-type="0"; x-mac-creator="0";
	name="timing-off-clarify.patch"
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment;
 filename="timing-off-clarify.patch"

diff --git a/doc/src/sgml/ref/explain.sgml b/doc/src/sgml/ref/explain.sgml
index 1f35a1d..627119b 100644
--- a/doc/src/sgml/ref/explain.sgml
+++ b/doc/src/sgml/ref/explain.sgml
@@ -186,7 +186,8 @@ ROLLBACK;
       The overhead of repeatedly reading the system clock can slow down the
       query significantly on some systems, so it may be useful to set this
       parameter to <literal>FALSE</literal> when only actual row counts, and not
-      exact times, are needed.
+      exact times, are needed.  Running time of the entire statement is always
+      collected, even when node level timing is turned off with this option. 
       This parameter may only be used when <literal>ANALYZE</literal> is also
       enabled.  It defaults to <literal>TRUE</literal>.
      </para>
diff --git a/src/backend/commands/explain.c b/src/backend/commands/explain.c
index 38ce0ef..90fc727 100644
--- a/src/backend/commands/explain.c
+++ b/src/backend/commands/explain.c
@@ -414,6 +414,10 @@ ExplainOnePlan(PlannedStmt *plannedstmt, IntoClause *into, ExplainState *es,
 	if (es->buffers)
 		instrument_option |= INSTRUMENT_BUFFERS;
 
+	/*
+	 * Always collect timing for the entire statement, even when
+	 * node level timing is off.
+	 */
 	INSTR_TIME_SET_CURRENT(starttime);
 
 	/*

--------------030400060101080808050004
Content-Type: text/plain
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
MIME-Version: 1.0


-- 
Sent via pgsql-docs mailing list (pgsql-docs@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-docs

--------------030400060101080808050004--