--- /dev/null
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
+<HTML>
+<HEAD>
+<TITLE>pg_store_plans</TITLE>
+<LINK
+REL="STYLESHEET"
+TYPE="text/css"
+HREF="stylesheet.css">
+<META
+HTTP-EQUIV="Content-Type"
+CONTENT="text/html; charset=ISO-8859-1">
+</HEAD>
+
+<BODY>
+<DIV CLASS="SECT1">
+<H1 CLASS="SECT1">
+<A NAME="PGSTOREPLANS">pg_store_plans</A>
+</H1>
+<P>The <TT CLASS="FILENAME">pg_store_plans</TT> module provides a
+ means for tracking execution plan statistics of all SQL statements
+ executed by a server.
+</P>
+
+<P>The module must be loaded by
+ adding <TT CLASS="LITERAL">pg_store_plans</TT> to
+ <A HREF="http://www.postgresql.org/docs/current/static/runtime-config-client.html#GUC-SHARED-PRELOAD-LIBRARIES">shared_preload_libraries</A> in
+ <TT CLASS="FILENAME">postgresql.conf</TT>, because it requires
+ additional shared memory. This means that a server restart is
+ required to add or remove the module.
+</P>
+<DIV CLASS="SECT2">
+<H2 CLASS="SECT2">1. The <TT CLASS="STRUCTNAME">pg_store_plans</TT>
+View</H2>
+
+<P>The statistics gathered by the module are available via a system
+ view named <TT CLASS="STRUCTNAME">pg_store_plans</TT>. This view
+ contains one row for each distinct set of database ID, user ID and
+ query ID. The columns of the view are described in
+ <A HREF="#PGSTOREPLANS-COLUMNS">Table 1</A>.
+</P>
+
+<DIV CLASS="TABLE">
+<A NAME="PGSTOREPLANS-COLUMNS"></A>
+<P><B>Table 1. <TT>pg_store_plans</TT> Columns</B>
+</P>
+<TABLE BORDER="1" CLASS="CALSTABLE" >
+<COL><COL><COL><COL>
+<THEAD><TR><TH>Name</TH><TH>Type</TH><TH>References</TH>
+ <TH>Description</TH></TR>
+</THEAD>
+<TBODY>
+<TR><TD><TT CLASS="STRUCTFIELD">userid</TT></TD>
+ <TD><TT CLASS="TYPE">oid</TT></TD>
+ <TD><A HREF="http://www.postgresql.org/docs/current/static/catalog-pg-authid.html">
+ <TT CLASS="STRUCTNAME">pg_authid</TT><TT CLASS="LITERAL"></A>.oid</TT></TD>
+ <TD>OID of user who executed the statement</TD></TR>
+<TR><TD><TT CLASS="STRUCTFIELD">dbid</TT></TD>
+ <TD><TT CLASS="TYPE">oid</TT></TD>
+ <TD><A HREF="http://www.postgresql.org/docs/current/static/catalog-pg-database.html">
+ <TT CLASS="STRUCTNAME">pg_database</TT></A><TT CLASS="LITERAL">.oid</TT></TD>
+ <TD>OID of database in which the statement was executed</TD></TR>
+<TR><TD><TT CLASS="STRUCTFIELD">queryid</TT></TD>
+ <TD><TT CLASS="TYPE">bigint</TT></TD>
+ <TD> </TD>
+ <TD>Internal hash code, computed from the statement's query string.</TD>
+ </TR>
+<TR><TD><TT CLASS="STRUCTFIELD">planid</TT></TD>
+ <TD><TT CLASS="TYPE">bigint</TT></TD>
+ <TD> </TD>
+ <TD>Internal hash code, computed from the statement's plan
+ representation.</TD>
+ </TR>
+<TR><TD><TT CLASS="STRUCTFIELD">queryid_stat_statements</TT></TD>
+ <TD><TT CLASS="TYPE">bigint</TT></TD>
+ <TD> </TD>
+ <TD>A copy of pg_stat_statements' query hash code. This is
+ available only when pg_stat_statements is installed.</TD>
+ </TR>
+<TR><TD><TT CLASS="STRUCTFIELD">plan</TT></TD>
+ <TD><TT CLASS="TYPE">text</TT></TD>
+ <TD> </TD>
+ <TD>Text of a representative plan. The format is specified by the
+ configuration
+ parameter <TT CLASS="VARNAME">pg_store_plans.plan_format.</TT></TD>
+ </TR>
+<TR><TD><TT CLASS="STRUCTFIELD">calls</TT></TD>
+ <TD><TT CLASS="TYPE">bigint</TT></TD>
+ <TD> </TD>
+ <TD>Number of times executed</TD>
+ </TR>
+<TR><TD><TT CLASS="STRUCTFIELD">total_time</TT></TD>
+ <TD><TT CLASS="TYPE">double precision</TT></TD>
+ <TD> </TD>
+ <TD>Total time spent in the statement using the plan, in milliseconds</TD>
+ </TR>
+<TR><TD><TT CLASS="STRUCTFIELD">rows</TT></TD>
+ <TD><TT CLASS="TYPE">bigint</TT></TD>
+ <TD> </TD>
+ <TD>Total number of rows retrieved or affected by the statement
+ using the plan</TD>
+ </TR>
+<TR><TD><TT CLASS="STRUCTFIELD">shared_blks_hit</TT></TD>
+ <TD><TT CLASS="TYPE">bigint</TT></TD>
+ <TD> </TD>
+ <TD>Total number of shared block cache hits by the statement using
+ the plan</TD>
+ </TR>
+<TR><TD><TT CLASS="STRUCTFIELD">shared_blks_read</TT></TD>
+ <TD><TT CLASS="TYPE">bigint</TT></TD>
+ <TD> </TD>
+ <TD>Total number of shared blocks read by the statement using the
+ plan</TD>
+ </TR>
+<TR><TD><TT CLASS="STRUCTFIELD">shared_blks_dirtied</TT></TD>
+ <TD><TT CLASS="TYPE">bigint</TT></TD>
+ <TD> </TD>
+ <TD>Total number of shared blocks dirtied by the statement using
+ the plan</TD>
+ </TR>
+<TR><TD><TT CLASS="STRUCTFIELD">shared_blks_written</TT></TD>
+ <TD><TT CLASS="TYPE">bigint</TT></TD>
+ <TD> </TD>
+ <TD>Total number of shared blocks written by the statement using
+ the plan</TD>
+ </TR>
+<TR><TD><TT CLASS="STRUCTFIELD">local_blks_hit</TT></TD>
+ <TD><TT CLASS="TYPE">bigint</TT></TD>
+ <TD> </TD>
+ <TD>Total number of local block cache hits by the statement using
+ the plan</TD>
+ </TR>
+<TR><TD><TT CLASS="STRUCTFIELD">local_blks_read</TT></TD>
+ <TD><TT CLASS="TYPE">bigint</TT></TD>
+ <TD> </TD>
+ <TD>Total number of local blocks read by the statement using the
+ plan</TD>
+ </TR>
+<TR><TD><TT CLASS="STRUCTFIELD">local_blks_dirtied</TT></TD>
+ <TD><TT CLASS="TYPE">bigint</TT></TD><TD> </TD>
+ <TD>Total number of local blocks dirtied by the statement using
+ the plan</TD>
+ </TR>
+<TR><TD><TT CLASS="STRUCTFIELD">local_blks_written</TT></TD>
+ <TD><TT CLASS="TYPE">bigint</TT></TD>
+ <TD> </TD>
+ <TD>Total number of local blocks written by the statement using
+ the plan</TD>
+ </TR>
+<TR><TD><TT CLASS="STRUCTFIELD">temp_blks_read</TT></TD>
+ <TD><TT CLASS="TYPE">bigint</TT></TD>
+ <TD> </TD>
+ <TD>Total number of temp blocks read by the statement using the
+ plan</TD>
+ </TR>
+<TR><TD><TT CLASS="STRUCTFIELD">temp_blks_written</TT></TD>
+ <TD><TT CLASS="TYPE">bigint</TT></TD>
+ <TD> </TD>
+ <TD>Total number of temp blocks written by the statement using the
+ plan</TD>
+ </TR>
+<TR><TD><TT CLASS="STRUCTFIELD">blk_read_time</TT></TD>
+ <TD><TT CLASS="TYPE">double precision</TT></TD>
+ <TD> </TD>
+
+ <TD>Total time the statement using the plan spent reading blocks, in milliseconds (if <A HREF="http://www.postgresql.org/docs/current/static/runtime-config-statistics.html#GUC-TRACK-IO-TIMING">track_io_timing</A> is enabled, otherwise zero)</TD>
+ </TR>
+<TR><TD><TT CLASS="STRUCTFIELD">blk_write_time</TT></TD>
+ <TD><TT CLASS="TYPE">double precision</TT></TD>
+ <TD> </TD>
+
+ <TD>Total time the statement using the plan spent writing blocks, in milliseconds (if <A HREF="http://www.postgresql.org/docs/current/static/runtime-config-statistics.html#GUC-TRACK-IO-TIMING">track_io_timing</A> is enabled, otherwise zero)</TD>
+ </TR>
+<TR><TD><TT CLASS="STRUCTFIELD">first_call</TT></TD>
+ <TD><TT CLASS="TYPE">timestamp with time zone</TT></TD>
+ <TD> </TD>
+ <TD>Timestamp for the least recently call of the query using this
+ plan.</TD>
+ </TR>
+<TR><TD><TT CLASS="STRUCTFIELD">last_call</TT></TD>
+ <TD><TT CLASS="TYPE">timestamp with time zone</TT></TD>
+ <TD> </TD>
+ <TD>Timestamp for the most recently call of the query using this
+ plan.</TD>
+ </TR>
+</TBODY>
+</TABLE>
+</DIV>
+
+<P>This view, and the functions <CODE CLASS="FUNCTION">pg_store_plans_reset
+ </CODE> and <CODE CLASS="FUNCTION">pg_store_plans</CODE> and other
+ auxiliary functions, are available only in databases where
+ the <TT CLASS="LITERAL">pg_store_plans</TT> is installed
+ by <TT CLASS="LITERAL">CREATE EXTENSION</TT>. However,
+ statistics are tracked across all databases of the server whenever
+ the <TT CLASS="FILENAME">pg_store_plans</TT> module is loaded onto
+ the server, regardless of presence of the view.
+</P>
+
+<P>For security reasons, non-superusers are not allowed to see the
+ plan representation, queryid or planid for the queries executed by
+ other users.
+</P>
+
+<P>
+
+<TT CLASS="VARNAME">queryid</TT> is calculated to identify the source
+query similary to <TT CLASS="structname">pg_stat_statements</TT> but
+in a different algorithm. <TT CLASS="VARNAME">plan</TT> is
+calculated in a similar way. Two plans are considered the same if they
+are seemingly equivalent except for the values of literal constants
+or fluctuating values such like costs or measured time. </P>
+
+<P> For PostgreSQL 9.4 or later, you can find the corresponding query
+for a <TT CLASS="structname">pg_store_plans</TT> entry
+in <TT CLASS="structname">pg_stat_statements</TT> by joining using
+<TT CLASS="structname">queryid_stat_statements</TT>. Otherwise it is
+identified by using <TT CLASS="VARNAME">queryid</TT>
+and <CODE CLASS="FUNCTION">pg_store_plans_hash_query</CODE >, like
+following.
+
+</P>
+<P>
+<PRE CLASS="SCREEN">SELECT s.query, p.plan FROM pg_store_plans p JOIN pg_stat_statements s ON (pg_store_plans_hash_query(s.query)) = p.queryid;</PRE>
+</P>
+
+<P> However plan id is calculated ignoring fluctuating values, the
+values for most recent execution are still displayed
+in <TT CLASS="STRUCTNAME">pg_store_plans.plan</TT>.
+</P>
+<P> In some cases, <TT CLASS="STRUCTNAME">pg_stat_statements</TT>
+ merges semantically equivalent queries which are considered
+ different by
+<TT CLASS="STRUCTNAME">pg_stat_statements</TT>. In the cases
+correspondent in <TT CLASS="STRUCTNAME">pg_stat_statements</TT> might
+not be found, but there is a small chance that this happenes. In
+contrast, there also is a small chance that some queries might be
+regarded as equivalent and merged into one entry
+in <TT CLASS="STRUCTNAME">pg_store_plans</TT> but differentiated
+in <TT CLASS="STRUCTNAME">pg_stat_statements</TT> mainly for utility
+statements.
+</P>
+
+<P><TT CLASS="STRUCTNAME">pg_store_plans</TT>
+ and <TT CLASS="STRUCTNAME">pg_stat_statements</TT> maintain thier
+ entries individually so there is certain unavoidable chance
+ especially for entries with low execution frequency that no
+ correspondent is found.
+</P>
+<P><TT CLASS="STRUCTFIELD">queryid_stat_statements</TT> has the same
+ restriction to <TT CLASS="STRUCTNAME">pg_stat_statements</TT> in
+ terms of stability. Although <TT CLASS="STRUCTFIELD">queryid</TT>
+ and <TT CLASS="STRUCTFIELD">planid</TT>
+ in <TT CLASS="STRUCTNAME">pg_store_plans</TT> doesn't have such a
+ restriction, assuming long-term stability is also discouraged.
+ </P>
+</DIV>
+<DIV CLASS="SECT2">
+<H2 CLASS="SECT2">
+<A NAME="Functions">2. Functions</A>
+</H2>
+<DIV CLASS="VARIABLELIST">
+<DL> <DT> <CODE CLASS="FUNCTION">pg_store_plans_reset() returns void</CODE>
+</DT>
+<DD>
+<P>
+ <CODE CLASS="FUNCTION">pg_store_plans_reset</CODE> discards all
+ statistics gathered so far
+ by <TT CLASS="FILENAME">pg_store_plans</TT>. By default, only
+ superusers can execute this function.
+ </P>
+</DD>
+<DT>
+<CODE CLASS="FUNCTION">pg_store_plans(showtext boolean) returns setof
+record</CODE>
+</DT>
+<DD>
+<P> The <TT CLASS="STRUCTNAME">pg_store_plans</TT> view is defined in
+ terms of a function also
+ named <CODE CLASS="FUNCTION">pg_store_plans</CODE >.
+</P>
+</DD>
+<DT>
+<CODE CLASS="FUNCTION">pg_store_hash_query(query text) returns oid</CODE>
+</DT>
+<DD>
+<P> This function calculates hash value of a query text. The same
+ algorithm is used to
+ calculate <TT CLASS="STRUCTFIELD">queryid</TT>
+ in <TT CLASS="STRUCTNAME">pg_store_plans</TT> so this function
+ is usable to join
+ with <TT CLASS="STRUCTNAME">pg_store_plans</TT>.
+ </P>
+</DD>
+<DT>
+<CODE CLASS="FUNCTION">pg_store_plans_textplan(query text) returns text</CODE>
+</DT>
+<DD>
+<P> This function generates a ordinary text representation from raw
+ representation of <TT CLASS="STRUCTFIELD">plan</TT>
+ in <TT CLASS="STRUCTNAME">pg_store_plans</TT>, which is shown
+ there when <TT CLASS="VARNAME">pg_store_plans.plan_formats</TT>
+ = 'raw'. Since the result plan text is generated from json
+ representation, it might be slightly different from what you
+ will get directly from 'EXPLAIN' commnand.
+ </P>
+</DD>
+<DT>
+<CODE CLASS="FUNCTION">pg_store_plans_jsonplan(query text) returns text</CODE>
+</DT>
+<DD>
+<P> This function infaltes a "short format json plan" or "raw format"
+ into normal json format. Short format json is internal format
+ for <TT CLASS="STRUCTFIELD">plan</TT>
+ in <TT CLASS="STRUCTNAME">pg_store_plans</TT>, which is shown
+ there when <TT CLASS="VARNAME">pg_store_plans.plan_formats</TT> =
+ 'raw'.
+ </P>
+</DD>
+<DT>
+<CODE CLASS="FUNCTION">pg_store_plans_xmlplan(query text) returns text</CODE>
+</DT>
+<DD>
+<P> This function generates a XML representation from raw
+ representation of <TT CLASS="STRUCTFIELD">plan</TT>
+ in <TT CLASS="STRUCTNAME">pg_store_plans</TT>, which is shown
+ there when <TT CLASS="VARNAME">pg_store_plans.plan_formats</TT>
+ = 'raw'.
+ </P>
+</DD>
+<DT>
+<CODE CLASS="FUNCTION">pg_store_plans_yamlplan(query text) returns text</CODE>
+</DT>
+<DD>
+<P> This function generates a YAML representation from raw
+ representation of <TT CLASS="STRUCTFIELD">plan</TT>
+ in <TT CLASS="STRUCTNAME">pg_store_plans</TT>, which is shown
+ there when <TT CLASS="VARNAME">pg_store_plans.plan_formats</TT>
+ = 'raw'.
+ </P>
+</DD>
+</DL>
+</DIV>
+</DIV>
+<DIV CLASS="SECT2">
+<H2 CLASS="SECT2">
+<ANAME="Config">3. Configuration Parameters</A>
+</H2>
+<P>
+</P>
+<DIV CLASS="VARIABLELIST">
+<DL>
+<DT>
+<TT CLASS="VARNAME">pg_store_plans.max</TT>
+ (<TT CLASS="TYPE">integer</TT>)</DT>
+<DD>
+<P> <TT CLASS="VARNAME">pg_store_plans.max</TT> is the maximum number
+of plans tracked by the module (i.e., the maximum number of rows in
+the <TT CLASS="STRUCTNAME">pg_store_plans</TT> view). If more
+distinct plans than that are observed, information about the
+least-executed plan is discarded. The default value is 1000. This
+parameter can only be set at server start.
+ </P>
+</DD>
+<DT>
+<TT CLASS="VARNAME">pg_store_plans.track</TT>
+ (<TT CLASS="TYPE">enum</TT>)
+</DT>
+<DD>
+<P> Similar to <TT CLASS="STRUCTNAME">pg_stat_statements</TT>,
+ <TT CLASS="VARNAME">pg_store_plans.track</TT> controls which
+ statements are counted by the module.
+ Specify <TT CLASS="LITERAL">top</TT> to track top-level
+ statements (those issued directly by
+ clients), <TT CLASS="LITERAL">all</TT> to also track nested
+ statements (such as statements invoked within functions),
+ or <TT CLASS="LITERAL">none</TT> to disable statement
+ statistics collection. The default value
+ is <TT CLASS="LITERAL">top</TT>. Only superusers can change
+ this setting.
+ </P>
+</DD>
+<DT>
+<TT CLASS="VARNAME">pg_store_plans.plan_format</TT>
+ (<TT CLASS="TYPE">enum</TT>)
+</DT>
+<DD>
+<P> <TT CLASS="VARNAME">pg_store_plans.plan_format</TT> controls the
+ format of <TT CLASS="STRUCTFIELD">plans</TT>
+ in <TT CLASS="STRUCTNAME">pg_store_plans</TT>. <TT CLASS="LITERAL">text</TT>
+ is the default value and to show in ordinary text
+ representation, <TT CLASS="LITERAL">json</TT>, <TT CLASS="LITERAL">xml</TT>
+ and <TT CLASS="LITERAL">yaml</TT> to show in corresponding format.
+ <TT CLASS="LITERAL">raw</TT> to get internal representation which
+ can be fed to <CODE CLASS="FUNCTION">pg_store_plans_*plan</CODE >
+ functions.
+</P>
+</DD>
+<DT>
+<TT CLASS="VARNAME">pg_store_plans.min_duration</TT>
+ (<TT CLASS="TYPE">integer</TT>)
+</DT>
+<DD>
+<P> <TT CLASS="VARNAME">pg_store_plans.min_duration</TT> is the
+ minumum statement execution time, in milliseconds, that will cause the
+ statement's plan to be logged. Setting this to zero (the default) logs
+ all plans. Only superuses can change this setting.
+</P>
+</DD>
+<DT>
+<TT CLASS="VARNAME">pg_store_plans.log_analyze</TT>
+ (<TT CLASS="TYPE">boolean</TT>)
+</DT>
+<DD>
+<P> <TT CLASS="VARNAME">pg_store_plans.log_analyze</TT>
+ causes <TT CLASS="COMMAND">EXPLAIN ANALYZE</TT> output, rather than
+ just <TT CLASS="COMMAND">EXPLAIN</TT> output, to be included
+ in <TT CLASS="STRUCTFIELD">plan</TT>. This parameter is off by
+ default.
+</P>
+</DD>
+<DT>
+<TT CLASS="VARNAME">pg_store_plans.log_buffers</TT>
+ (<TT CLASS="TYPE">boolean</TT>)
+</DT>
+<DD>
+<P> <TT CLASS="VARNAME">pg_store_plans.log_buffers</TT>
+ causes <TT CLASS="COMMAND">EXPLAIN (ANALYZE, BUFFERS)</TT> output,
+ rather than just <TT CLASS="COMMAND">EXPLAIN</TT> output, to be
+ included in <TT CLASS="STRUCTFIELD">plan</TT>. This parameter is off
+ by default.
+</P>
+</DD>
+<DT>
+<TT CLASS="VARNAME">pg_store_plans.log_timing</TT>
+ (<TT CLASS="TYPE">boolean</TT>)</DT>
+<DD>
+<P> Setting <TT CLASS="VARNAME">pg_store_plans.log_timing</TT> to
+ false disables to record actual timings. 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 FALSE
+ when only actual row counts, and not exact execution times for each
+ execution nodes, are needed. Run time of the entire statement is
+ always measured
+ when <TT CLASS="VARNAME">pg_store_plans.log_analyze</TT> is
+ TRUE. It defaults to TRUE.
+</P>
+</DD>
+<DT>
+<TT CLASS="VARNAME">pg_store_plans.log_triggers</TT
+> (<TT CLASS="TYPE">boolean</TT
+>)</DT>
+<DD>
+<P> <TT CLASS="VARNAME">pg_store_plans.log_triggers</TT> causes
+ trigger execution statistics to be included in recoreded plans. This
+ parameter has no effect
+ unless <TT CLASS="VARNAME">pg_store_plans.log_analyze</TT> is
+ turned on.
+</P>
+</DD>
+<DT>
+<TT CLASS="VARNAME">pg_store_plans.verbose</TT>
+ (<TT CLASS="TYPE">boolean</TT>)
+</DT>
+<DD>
+<P> <TT CLASS="VARNAME">pg_store_plans.verbose</TT>
+ causes <TT CLASS="COMMAND">EXPLAIN VERBOSE</TT> output, rather than
+ just <TT CLASS="COMMAND">EXPLAIN</TT> output, to be included
+ in <TT CLASS="STRUCTFIELD">plan</TT>. This parameter is off by
+ default.
+</P>
+</DD>
+<DT>
+<TT CLASS="VARNAME">pg_store_plans.save</TT>
+ (<TT CLASS="TYPE">boolean</TT>)
+</DT>
+<DD>
+<P> <TT CLASS="VARNAME">pg_store_plans.save</TT> specifies whether to
+ save plan statistics across server shutdowns. If it
+ is <TT CLASS="LITERAL">off</TT> then statistics are not saved at
+ shutdown nor reloaded at server start. The default value
+ is <TT CLASS="LITERAL">on</TT>. This parameter can only be set in
+ the <TT CLASS="FILENAME">postgresql.conf</TT> file or on the server
+ command line.
+</P>
+</DD>
+</DL>
+</DIV>
+<P> The module requires additional shared memory proportional to
+ <TT CLASS="VARNAME">pg_store_plans.max</TT>. Note that this memory
+ is consumed whenever the module is loaded, even if
+ <TT CLASS="VARNAME">pg_store_plans.track</TT> is set
+ to <TT CLASS="LITERAL">none</TT>.
+</P>
+<P> These parameters must be set in
+ <TT CLASS="FILENAME">postgresql.conf</TT>. Typical usage might be:
+</P><PRE CLASS="PROGRAMLISTING"># postgresql.conf
+shared_preload_libraries = 'pg_store_plans, pg_stat_statements'
+pg_store_plans.max = 10000
+pg_store_plans.track = all</PRE>
+<P>
+</P>
+</DIV>
+<DIV CLASS="SECT2">
+<H2 CLASS="SECT2">
+<A NAME="Sample">4. Sample Output</A>
+</H2>
+<PRE CLASS="SCREEN">(postgresql.conf has following settings)
+shared_preload_libraries = 'pg_store_plans,pg_stat_statements'
+pg_store_plans.log_analyze = true
+pg_store_plans.log_timing = false
+
+bench=# SELECT pg_store_plans_reset();
+
+$ pgbench -i bench
+$ pgbench -c10 -t3000 bench
+
+bench=# \x
+bench=# SELECT s.query, p.plan,
+ p.calls as "plan calls", s.calls as "stmt calls",
+ p.total_time / p.calls as "time/call", p.first_call, p.last_call
+ FROM pg_stat_statements s
+ JOIN pg_store_plans p ON
+ (p.queryid = pg_store_plans_hash_query(s.query) and p.calls < s.calls)
+ ORDER BY query ASC, "time/call" DESC;
+-[ RECORD 1 ]----------------------------------------------------------------------------------------------------------------------------
+query | UPDATE pgbench_branches SET bbalance = bbalance + ? WHERE bid = ?;
+plan | Update on pgbench_branches (cost=0.00..8.01 rows=1 width=370) (actual rows=0 loops=1)
+ | -> Seq Scan on pgbench_branches (cost=0.00..8.01 rows=1 width=370) (actual rows=1 loops=1)
+ | Filter: (bid = 1)
+plan calls | 15583
+stmt calls | 30000
+time/call | 40.096513957518
+first_call | 2014-04-25 14:29:17.163924+09
+last_call | 2014-04-25 14:31:29.421635+09
+-[ RECORD 2 ]----------------------------------------------------------------------------------------------------------------------------
+query | UPDATE pgbench_branches SET bbalance = bbalance + ? WHERE bid = ?;
+plan | Update on pgbench_branches (cost=0.12..8.14 rows=1 width=370) (actual rows=0 loops=1)
+ | -> Index Scan using pgbench_branches_pkey on pgbench_branches (cost=0.12..8.14 rows=1 width=370) (actual rows=1 loops=1)
+ | Index Cond: (bid = 1)
+plan calls | 14417
+stmt calls | 30000
+time/call | 39.1920771311645
+first_call | 2014-04-25 14:31:29.288913+09
+last_call | 2014-04-25 14:33:31.287061+09
+-[ RECORD 3 ]----------------------------------------------------------------------------------------------------------------------------
+query | UPDATE pgbench_tellers SET tbalance = tbalance + ? WHERE tid = ?;
+plan | Update on pgbench_tellers (cost=0.14..8.16 rows=1 width=358) (actual rows=0 loops=1)
+ | -> Index Scan using pgbench_tellers_pkey on pgbench_tellers (cost=0.14..8.16 rows=1 width=358) (actual rows=1 loops=1)
+ | Index Cond: (tid = 7)
+plan calls | 4
+stmt calls | 30000
+time/call | 87.0435
+first_call | 2014-04-25 14:30:37.850293+09
+last_call | 2014-04-25 14:32:38.083977+09
+-[ RECORD 4 ]----------------------------------------------------------------------------------------------------------------------------
+query | UPDATE pgbench_tellers SET tbalance = tbalance + ? WHERE tid = ?;
+plan | Update on pgbench_tellers (cost=4.14..8.16 rows=1 width=358) (actual rows=0 loops=1)
+ | -> Bitmap Heap Scan on pgbench_tellers (cost=4.14..8.16 rows=1 width=358) (actual rows=1 loops=1)
+ | Recheck Cond: (tid = 10)
+ | -> Bitmap Index Scan using pgbench_tellers_pkey (cost=0.00..4.14 rows=1 width=0) (actual rows=1 loops=1)
+ | Index Cond: (tid = 10)
+plan calls | 29996
+stmt calls | 30000
+time/call | 33.6455953793834
+first_call | 2014-04-25 14:29:17.162871+09
+last_call | 2014-04-25 14:33:31.28646+09</PRE>
+</DIV>
+</DIV>
+<HR>
+</BODY>
+</HTML>
OSDN Git Service
