codekingpro/portable-devtools
115k
1<?xml version="1.0" encoding="UTF-8" standalone="no"?>2<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /><title>20.7. Query Planning</title><link rel="stylesheet" type="text/css" href="stylesheet.css" /><link rev="made" href="pgsql-docs@lists.postgresql.org" /><meta name="generator" content="DocBook XSL Stylesheets Vsnapshot" /><link rel="prev" href="runtime-config-replication.html" title="20.6. Replication" /><link rel="next" href="runtime-config-logging.html" title="20.8. Error Reporting and Logging" /></head><body id="docContent" class="container-fluid col-10"><div class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">20.7. Query Planning</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="runtime-config-replication.html" title="20.6. Replication">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="runtime-config.html" title="Chapter 20. Server Configuration">Up</a></td><th width="60%" align="center">Chapter 20. Server Configuration</th><td width="10%" align="right"><a accesskey="h" href="index.html" title="PostgreSQL 16.3 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="runtime-config-logging.html" title="20.8. Error Reporting and Logging">Next</a></td></tr></table><hr /></div><div class="sect1" id="RUNTIME-CONFIG-QUERY"><div class="titlepage"><div><div><h2 class="title" style="clear: both">20.7. Query Planning <a href="#RUNTIME-CONFIG-QUERY" class="id_link">#</a></h2></div></div></div><div class="toc"><dl class="toc"><dt><span class="sect2"><a href="runtime-config-query.html#RUNTIME-CONFIG-QUERY-ENABLE">20.7.1. Planner Method Configuration</a></span></dt><dt><span class="sect2"><a href="runtime-config-query.html#RUNTIME-CONFIG-QUERY-CONSTANTS">20.7.2. Planner Cost Constants</a></span></dt><dt><span class="sect2"><a href="runtime-config-query.html#RUNTIME-CONFIG-QUERY-GEQO">20.7.3. Genetic Query Optimizer</a></span></dt><dt><span class="sect2"><a href="runtime-config-query.html#RUNTIME-CONFIG-QUERY-OTHER">20.7.4. Other Planner Options</a></span></dt></dl></div><div class="sect2" id="RUNTIME-CONFIG-QUERY-ENABLE"><div class="titlepage"><div><div><h3 class="title">20.7.1. Planner Method Configuration <a href="#RUNTIME-CONFIG-QUERY-ENABLE" class="id_link">#</a></h3></div></div></div><p>3 These configuration parameters provide a crude method of4 influencing the query plans chosen by the query optimizer. If5 the default plan chosen by the optimizer for a particular query6 is not optimal, a <span class="emphasis"><em>temporary</em></span> solution is to use one7 of these configuration parameters to force the optimizer to8 choose a different plan.9 Better ways to improve the quality of the10 plans chosen by the optimizer include adjusting the planner cost11 constants (see <a class="xref" href="runtime-config-query.html#RUNTIME-CONFIG-QUERY-CONSTANTS" title="20.7.2. Planner Cost Constants">Section 20.7.2</a>),12 running <a class="link" href="sql-analyze.html" title="ANALYZE"><code class="command">ANALYZE</code></a> manually, increasing13 the value of the <a class="xref" href="runtime-config-query.html#GUC-DEFAULT-STATISTICS-TARGET">default_statistics_target</a> configuration parameter,14 and increasing the amount of statistics collected for15 specific columns using <code class="command">ALTER TABLE SET16 STATISTICS</code>.17 </p><div class="variablelist"><dl class="variablelist"><dt id="GUC-ENABLE-ASYNC-APPEND"><span class="term"><code class="varname">enable_async_append</code> (<code class="type">boolean</code>)18 <a id="id-1.6.7.10.2.3.1.1.3" class="indexterm"></a>19 </span> <a href="#GUC-ENABLE-ASYNC-APPEND" class="id_link">#</a></dt><dd><p>20 Enables or disables the query planner's use of async-aware21 append plan types. The default is <code class="literal">on</code>.22 </p></dd><dt id="GUC-ENABLE-BITMAPSCAN"><span class="term"><code class="varname">enable_bitmapscan</code> (<code class="type">boolean</code>)23 <a id="id-1.6.7.10.2.3.2.1.3" class="indexterm"></a>24 <a id="id-1.6.7.10.2.3.2.1.4" class="indexterm"></a>25 </span> <a href="#GUC-ENABLE-BITMAPSCAN" class="id_link">#</a></dt><dd><p>26 Enables or disables the query planner's use of bitmap-scan plan27 types. The default is <code class="literal">on</code>.28 </p></dd><dt id="GUC-ENABLE-GATHERMERGE"><span class="term"><code class="varname">enable_gathermerge</code> (<code class="type">boolean</code>)29 <a id="id-1.6.7.10.2.3.3.1.3" class="indexterm"></a>30 </span> <a href="#GUC-ENABLE-GATHERMERGE" class="id_link">#</a></dt><dd><p>31 Enables or disables the query planner's use of gather32 merge plan types. The default is <code class="literal">on</code>.33 </p></dd><dt id="GUC-ENABLE-HASHAGG"><span class="term"><code class="varname">enable_hashagg</code> (<code class="type">boolean</code>)34 <a id="id-1.6.7.10.2.3.4.1.3" class="indexterm"></a>35 </span> <a href="#GUC-ENABLE-HASHAGG" class="id_link">#</a></dt><dd><p>36 Enables or disables the query planner's use of hashed37 aggregation plan types. The default is <code class="literal">on</code>.38 </p></dd><dt id="GUC-ENABLE-HASHJOIN"><span class="term"><code class="varname">enable_hashjoin</code> (<code class="type">boolean</code>)39 <a id="id-1.6.7.10.2.3.5.1.3" class="indexterm"></a>40 </span> <a href="#GUC-ENABLE-HASHJOIN" class="id_link">#</a></dt><dd><p>41 Enables or disables the query planner's use of hash-join plan42 types. The default is <code class="literal">on</code>.43 </p></dd><dt id="GUC-ENABLE-INCREMENTAL-SORT"><span class="term"><code class="varname">enable_incremental_sort</code> (<code class="type">boolean</code>)44 <a id="id-1.6.7.10.2.3.6.1.3" class="indexterm"></a>45 </span> <a href="#GUC-ENABLE-INCREMENTAL-SORT" class="id_link">#</a></dt><dd><p>46 Enables or disables the query planner's use of incremental sort steps.47 The default is <code class="literal">on</code>.48 </p></dd><dt id="GUC-ENABLE-INDEXSCAN"><span class="term"><code class="varname">enable_indexscan</code> (<code class="type">boolean</code>)49 <a id="id-1.6.7.10.2.3.7.1.3" class="indexterm"></a>50 <a id="id-1.6.7.10.2.3.7.1.4" class="indexterm"></a>51 </span> <a href="#GUC-ENABLE-INDEXSCAN" class="id_link">#</a></dt><dd><p>52 Enables or disables the query planner's use of index-scan plan53 types. The default is <code class="literal">on</code>.54 </p></dd><dt id="GUC-ENABLE-INDEXONLYSCAN"><span class="term"><code class="varname">enable_indexonlyscan</code> (<code class="type">boolean</code>)55 <a id="id-1.6.7.10.2.3.8.1.3" class="indexterm"></a>56 </span> <a href="#GUC-ENABLE-INDEXONLYSCAN" class="id_link">#</a></dt><dd><p>57 Enables or disables the query planner's use of index-only-scan plan58 types (see <a class="xref" href="indexes-index-only-scans.html" title="11.9. Index-Only Scans and Covering Indexes">Section 11.9</a>).59 The default is <code class="literal">on</code>.60 </p></dd><dt id="GUC-ENABLE-MATERIAL"><span class="term"><code class="varname">enable_material</code> (<code class="type">boolean</code>)61 <a id="id-1.6.7.10.2.3.9.1.3" class="indexterm"></a>62 </span> <a href="#GUC-ENABLE-MATERIAL" class="id_link">#</a></dt><dd><p>63 Enables or disables the query planner's use of materialization.64 It is impossible to suppress materialization entirely,65 but turning this variable off prevents the planner from inserting66 materialize nodes except in cases where it is required for correctness.67 The default is <code class="literal">on</code>.68 </p></dd><dt id="GUC-ENABLE-MEMOIZE"><span class="term"><code class="varname">enable_memoize</code> (<code class="type">boolean</code>)69 <a id="id-1.6.7.10.2.3.10.1.3" class="indexterm"></a>70 </span> <a href="#GUC-ENABLE-MEMOIZE" class="id_link">#</a></dt><dd><p>71 Enables or disables the query planner's use of memoize plans for72 caching results from parameterized scans inside nested-loop joins.73 This plan type allows scans to the underlying plans to be skipped when74 the results for the current parameters are already in the cache. Less75 commonly looked up results may be evicted from the cache when more76 space is required for new entries. The default is77 <code class="literal">on</code>.78 </p></dd><dt id="GUC-ENABLE-MERGEJOIN"><span class="term"><code class="varname">enable_mergejoin</code> (<code class="type">boolean</code>)79 <a id="id-1.6.7.10.2.3.11.1.3" class="indexterm"></a>80 </span> <a href="#GUC-ENABLE-MERGEJOIN" class="id_link">#</a></dt><dd><p>81 Enables or disables the query planner's use of merge-join plan82 types. The default is <code class="literal">on</code>.83 </p></dd><dt id="GUC-ENABLE-NESTLOOP"><span class="term"><code class="varname">enable_nestloop</code> (<code class="type">boolean</code>)84 <a id="id-1.6.7.10.2.3.12.1.3" class="indexterm"></a>85 </span> <a href="#GUC-ENABLE-NESTLOOP" class="id_link">#</a></dt><dd><p>86 Enables or disables the query planner's use of nested-loop join87 plans. It is impossible to suppress nested-loop joins entirely,88 but turning this variable off discourages the planner from using89 one if there are other methods available. The default is90 <code class="literal">on</code>.91 </p></dd><dt id="GUC-ENABLE-PARALLEL-APPEND"><span class="term"><code class="varname">enable_parallel_append</code> (<code class="type">boolean</code>)92 <a id="id-1.6.7.10.2.3.13.1.3" class="indexterm"></a>93 </span> <a href="#GUC-ENABLE-PARALLEL-APPEND" class="id_link">#</a></dt><dd><p>94 Enables or disables the query planner's use of parallel-aware95 append plan types. The default is <code class="literal">on</code>.96 </p></dd><dt id="GUC-ENABLE-PARALLEL-HASH"><span class="term"><code class="varname">enable_parallel_hash</code> (<code class="type">boolean</code>)97 <a id="id-1.6.7.10.2.3.14.1.3" class="indexterm"></a>98 </span> <a href="#GUC-ENABLE-PARALLEL-HASH" class="id_link">#</a></dt><dd><p>99 Enables or disables the query planner's use of hash-join plan100 types with parallel hash. Has no effect if hash-join plans are not101 also enabled. The default is <code class="literal">on</code>.102 </p></dd><dt id="GUC-ENABLE-PARTITION-PRUNING"><span class="term"><code class="varname">enable_partition_pruning</code> (<code class="type">boolean</code>)103 <a id="id-1.6.7.10.2.3.15.1.3" class="indexterm"></a>104 </span> <a href="#GUC-ENABLE-PARTITION-PRUNING" class="id_link">#</a></dt><dd><p>105 Enables or disables the query planner's ability to eliminate a106 partitioned table's partitions from query plans. This also controls107 the planner's ability to generate query plans which allow the query108 executor to remove (ignore) partitions during query execution. The109 default is <code class="literal">on</code>.110 See <a class="xref" href="ddl-partitioning.html#DDL-PARTITION-PRUNING" title="5.11.4. Partition Pruning">Section 5.11.4</a> for details.111 </p></dd><dt id="GUC-ENABLE-PARTITIONWISE-JOIN"><span class="term"><code class="varname">enable_partitionwise_join</code> (<code class="type">boolean</code>)112 <a id="id-1.6.7.10.2.3.16.1.3" class="indexterm"></a>113 </span> <a href="#GUC-ENABLE-PARTITIONWISE-JOIN" class="id_link">#</a></dt><dd><p>114 Enables or disables the query planner's use of partitionwise join,115 which allows a join between partitioned tables to be performed by116 joining the matching partitions. Partitionwise join currently applies117 only when the join conditions include all the partition keys, which118 must be of the same data type and have one-to-one matching sets of119 child partitions. Because partitionwise join planning can use120 significantly more CPU time and memory during planning, the default is121 <code class="literal">off</code>.122 </p></dd><dt id="GUC-ENABLE-PARTITIONWISE-AGGREGATE"><span class="term"><code class="varname">enable_partitionwise_aggregate</code> (<code class="type">boolean</code>)123 <a id="id-1.6.7.10.2.3.17.1.3" class="indexterm"></a>124 </span> <a href="#GUC-ENABLE-PARTITIONWISE-AGGREGATE" class="id_link">#</a></dt><dd><p>125 Enables or disables the query planner's use of partitionwise grouping126 or aggregation, which allows grouping or aggregation on partitioned127 tables to be performed separately for each partition. If the128 <code class="literal">GROUP BY</code> clause does not include the partition129 keys, only partial aggregation can be performed on a per-partition130 basis, and finalization must be performed later. Because131 partitionwise grouping or aggregation can use significantly more CPU132 time and memory during planning, the default is133 <code class="literal">off</code>.134 </p></dd><dt id="GUC-ENABLE-PRESORTED-AGGREGATE"><span class="term"><code class="varname">enable_presorted_aggregate</code> (<code class="type">boolean</code>)135 <a id="id-1.6.7.10.2.3.18.1.3" class="indexterm"></a>136 </span> <a href="#GUC-ENABLE-PRESORTED-AGGREGATE" class="id_link">#</a></dt><dd><p>137 Controls if the query planner will produce a plan which will provide138 rows which are presorted in the order required for the query's139 <code class="literal">ORDER BY</code> / <code class="literal">DISTINCT</code> aggregate140 functions. When disabled, the query planner will produce a plan which141 will always require the executor to perform a sort before performing142 aggregation of each aggregate function containing an143 <code class="literal">ORDER BY</code> or <code class="literal">DISTINCT</code> clause.144 When enabled, the planner will try to produce a more efficient plan145 which provides input to the aggregate functions which is presorted in146 the order they require for aggregation. The default value is147 <code class="literal">on</code>.148 </p></dd><dt id="GUC-ENABLE-SEQSCAN"><span class="term"><code class="varname">enable_seqscan</code> (<code class="type">boolean</code>)149 <a id="id-1.6.7.10.2.3.19.1.3" class="indexterm"></a>150 <a id="id-1.6.7.10.2.3.19.1.4" class="indexterm"></a>151 </span> <a href="#GUC-ENABLE-SEQSCAN" class="id_link">#</a></dt><dd><p>152 Enables or disables the query planner's use of sequential scan153 plan types. It is impossible to suppress sequential scans154 entirely, but turning this variable off discourages the planner155 from using one if there are other methods available. The156 default is <code class="literal">on</code>.157 </p></dd><dt id="GUC-ENABLE-SORT"><span class="term"><code class="varname">enable_sort</code> (<code class="type">boolean</code>)158 <a id="id-1.6.7.10.2.3.20.1.3" class="indexterm"></a>159 </span> <a href="#GUC-ENABLE-SORT" class="id_link">#</a></dt><dd><p>160 Enables or disables the query planner's use of explicit sort161 steps. It is impossible to suppress explicit sorts entirely,162 but turning this variable off discourages the planner from163 using one if there are other methods available. The default164 is <code class="literal">on</code>.165 </p></dd><dt id="GUC-ENABLE-TIDSCAN"><span class="term"><code class="varname">enable_tidscan</code> (<code class="type">boolean</code>)166 <a id="id-1.6.7.10.2.3.21.1.3" class="indexterm"></a>167 </span> <a href="#GUC-ENABLE-TIDSCAN" class="id_link">#</a></dt><dd><p>168 Enables or disables the query planner's use of <acronym class="acronym">TID</acronym>169 scan plan types. The default is <code class="literal">on</code>.170 </p></dd></dl></div></div><div class="sect2" id="RUNTIME-CONFIG-QUERY-CONSTANTS"><div class="titlepage"><div><div><h3 class="title">20.7.2. Planner Cost Constants <a href="#RUNTIME-CONFIG-QUERY-CONSTANTS" class="id_link">#</a></h3></div></div></div><p>171 The <em class="firstterm">cost</em> variables described in this section are measured172 on an arbitrary scale. Only their relative values matter, hence173 scaling them all up or down by the same factor will result in no change174 in the planner's choices. By default, these cost variables are based on175 the cost of sequential page fetches; that is,176 <code class="varname">seq_page_cost</code> is conventionally set to <code class="literal">1.0</code>177 and the other cost variables are set with reference to that. But178 you can use a different scale if you prefer, such as actual execution179 times in milliseconds on a particular machine.180 </p><div class="note"><h3 class="title">Note</h3><p>181 Unfortunately, there is no well-defined method for determining ideal182 values for the cost variables. They are best treated as averages over183 the entire mix of queries that a particular installation will receive. This184 means that changing them on the basis of just a few experiments is very185 risky.186 </p></div><div class="variablelist"><dl class="variablelist"><dt id="GUC-SEQ-PAGE-COST"><span class="term"><code class="varname">seq_page_cost</code> (<code class="type">floating point</code>)187 <a id="id-1.6.7.10.3.4.1.1.3" class="indexterm"></a>188 </span> <a href="#GUC-SEQ-PAGE-COST" class="id_link">#</a></dt><dd><p>189 Sets the planner's estimate of the cost of a disk page fetch190 that is part of a series of sequential fetches. The default is 1.0.191 This value can be overridden for tables and indexes in a particular192 tablespace by setting the tablespace parameter of the same name193 (see <a class="xref" href="sql-altertablespace.html" title="ALTER TABLESPACE"><span class="refentrytitle">ALTER TABLESPACE</span></a>).194 </p></dd><dt id="GUC-RANDOM-PAGE-COST"><span class="term"><code class="varname">random_page_cost</code> (<code class="type">floating point</code>)195 <a id="id-1.6.7.10.3.4.2.1.3" class="indexterm"></a>196 </span> <a href="#GUC-RANDOM-PAGE-COST" class="id_link">#</a></dt><dd><p>197 Sets the planner's estimate of the cost of a198 non-sequentially-fetched disk page. The default is 4.0.199 This value can be overridden for tables and indexes in a particular200 tablespace by setting the tablespace parameter of the same name201 (see <a class="xref" href="sql-altertablespace.html" title="ALTER TABLESPACE"><span class="refentrytitle">ALTER TABLESPACE</span></a>).202 </p><p>203 Reducing this value relative to <code class="varname">seq_page_cost</code>204 will cause the system to prefer index scans; raising it will205 make index scans look relatively more expensive. You can raise206 or lower both values together to change the importance of disk I/O207 costs relative to CPU costs, which are described by the following208 parameters.209 </p><p>210 Random access to mechanical disk storage is normally much more expensive211 than four times sequential access. However, a lower default is used212 (4.0) because the majority of random accesses to disk, such as indexed213 reads, are assumed to be in cache. The default value can be thought of214 as modeling random access as 40 times slower than sequential, while215 expecting 90% of random reads to be cached.216 </p><p>217 If you believe a 90% cache rate is an incorrect assumption218 for your workload, you can increase random_page_cost to better219 reflect the true cost of random storage reads. Correspondingly,220 if your data is likely to be completely in cache, such as when221 the database is smaller than the total server memory, decreasing222 random_page_cost can be appropriate. Storage that has a low random223 read cost relative to sequential, e.g., solid-state drives, might224 also be better modeled with a lower value for random_page_cost,225 e.g., <code class="literal">1.1</code>.226 </p><div class="tip"><h3 class="title">Tip</h3><p>227 Although the system will let you set <code class="varname">random_page_cost</code> to228 less than <code class="varname">seq_page_cost</code>, it is not physically sensible229 to do so. However, setting them equal makes sense if the database230 is entirely cached in RAM, since in that case there is no penalty231 for touching pages out of sequence. Also, in a heavily-cached232 database you should lower both values relative to the CPU parameters,233 since the cost of fetching a page already in RAM is much smaller234 than it would normally be.235 </p></div></dd><dt id="GUC-CPU-TUPLE-COST"><span class="term"><code class="varname">cpu_tuple_cost</code> (<code class="type">floating point</code>)236 <a id="id-1.6.7.10.3.4.3.1.3" class="indexterm"></a>237 </span> <a href="#GUC-CPU-TUPLE-COST" class="id_link">#</a></dt><dd><p>238 Sets the planner's estimate of the cost of processing239 each row during a query.240 The default is 0.01.241 </p></dd><dt id="GUC-CPU-INDEX-TUPLE-COST"><span class="term"><code class="varname">cpu_index_tuple_cost</code> (<code class="type">floating point</code>)242 <a id="id-1.6.7.10.3.4.4.1.3" class="indexterm"></a>243 </span> <a href="#GUC-CPU-INDEX-TUPLE-COST" class="id_link">#</a></dt><dd><p>244 Sets the planner's estimate of the cost of processing245 each index entry during an index scan.246 The default is 0.005.247 </p></dd><dt id="GUC-CPU-OPERATOR-COST"><span class="term"><code class="varname">cpu_operator_cost</code> (<code class="type">floating point</code>)248 <a id="id-1.6.7.10.3.4.5.1.3" class="indexterm"></a>249 </span> <a href="#GUC-CPU-OPERATOR-COST" class="id_link">#</a></dt><dd><p>250 Sets the planner's estimate of the cost of processing each251 operator or function executed during a query.252 The default is 0.0025.253 </p></dd><dt id="GUC-PARALLEL-SETUP-COST"><span class="term"><code class="varname">parallel_setup_cost</code> (<code class="type">floating point</code>)254 <a id="id-1.6.7.10.3.4.6.1.3" class="indexterm"></a>255 </span> <a href="#GUC-PARALLEL-SETUP-COST" class="id_link">#</a></dt><dd><p>256 Sets the planner's estimate of the cost of launching parallel worker257 processes.258 The default is 1000.259 </p></dd><dt id="GUC-PARALLEL-TUPLE-COST"><span class="term"><code class="varname">parallel_tuple_cost</code> (<code class="type">floating point</code>)260 <a id="id-1.6.7.10.3.4.7.1.3" class="indexterm"></a>261 </span> <a href="#GUC-PARALLEL-TUPLE-COST" class="id_link">#</a></dt><dd><p>262 Sets the planner's estimate of the cost of transferring one tuple263 from a parallel worker process to another process.264 The default is 0.1.265 </p></dd><dt id="GUC-MIN-PARALLEL-TABLE-SCAN-SIZE"><span class="term"><code class="varname">min_parallel_table_scan_size</code> (<code class="type">integer</code>)266 <a id="id-1.6.7.10.3.4.8.1.3" class="indexterm"></a>267 </span> <a href="#GUC-MIN-PARALLEL-TABLE-SCAN-SIZE" class="id_link">#</a></dt><dd><p>268 Sets the minimum amount of table data that must be scanned in order269 for a parallel scan to be considered. For a parallel sequential scan,270 the amount of table data scanned is always equal to the size of the271 table, but when indexes are used the amount of table data272 scanned will normally be less.273 If this value is specified without units, it is taken as blocks,274 that is <code class="symbol">BLCKSZ</code> bytes, typically 8kB.275 The default is 8 megabytes (<code class="literal">8MB</code>).276 </p></dd><dt id="GUC-MIN-PARALLEL-INDEX-SCAN-SIZE"><span class="term"><code class="varname">min_parallel_index_scan_size</code> (<code class="type">integer</code>)277 <a id="id-1.6.7.10.3.4.9.1.3" class="indexterm"></a>278 </span> <a href="#GUC-MIN-PARALLEL-INDEX-SCAN-SIZE" class="id_link">#</a></dt><dd><p>279 Sets the minimum amount of index data that must be scanned in order280 for a parallel scan to be considered. Note that a parallel index scan281 typically won't touch the entire index; it is the number of pages282 which the planner believes will actually be touched by the scan which283 is relevant. This parameter is also used to decide whether a284 particular index can participate in a parallel vacuum. See285 <a class="xref" href="sql-vacuum.html" title="VACUUM"><span class="refentrytitle">VACUUM</span></a>.286 If this value is specified without units, it is taken as blocks,287 that is <code class="symbol">BLCKSZ</code> bytes, typically 8kB.288 The default is 512 kilobytes (<code class="literal">512kB</code>).289 </p></dd><dt id="GUC-EFFECTIVE-CACHE-SIZE"><span class="term"><code class="varname">effective_cache_size</code> (<code class="type">integer</code>)290 <a id="id-1.6.7.10.3.4.10.1.3" class="indexterm"></a>291 </span> <a href="#GUC-EFFECTIVE-CACHE-SIZE" class="id_link">#</a></dt><dd><p>292 Sets the planner's assumption about the effective size of the293 disk cache that is available to a single query. This is294 factored into estimates of the cost of using an index; a295 higher value makes it more likely index scans will be used, a296 lower value makes it more likely sequential scans will be297 used. When setting this parameter you should consider both298 <span class="productname">PostgreSQL</span>'s shared buffers and the299 portion of the kernel's disk cache that will be used for300 <span class="productname">PostgreSQL</span> data files, though some301 data might exist in both places. Also, take302 into account the expected number of concurrent queries on different303 tables, since they will have to share the available304 space. This parameter has no effect on the size of shared305 memory allocated by <span class="productname">PostgreSQL</span>, nor306 does it reserve kernel disk cache; it is used only for estimation307 purposes. The system also does not assume data remains in308 the disk cache between queries.309 If this value is specified without units, it is taken as blocks,310 that is <code class="symbol">BLCKSZ</code> bytes, typically 8kB.311 The default is 4 gigabytes (<code class="literal">4GB</code>).312 (If <code class="symbol">BLCKSZ</code> is not 8kB, the default value scales313 proportionally to it.)314 </p></dd><dt id="GUC-JIT-ABOVE-COST"><span class="term"><code class="varname">jit_above_cost</code> (<code class="type">floating point</code>)315 <a id="id-1.6.7.10.3.4.11.1.3" class="indexterm"></a>316 </span> <a href="#GUC-JIT-ABOVE-COST" class="id_link">#</a></dt><dd><p>317 Sets the query cost above which JIT compilation is activated, if318 enabled (see <a class="xref" href="jit.html" title="Chapter 32. Just-in-Time Compilation (JIT)">Chapter 32</a>).319 Performing <acronym class="acronym">JIT</acronym> costs planning time but can320 accelerate query execution.321 Setting this to <code class="literal">-1</code> disables JIT compilation.322 The default is <code class="literal">100000</code>.323 </p></dd><dt id="GUC-JIT-INLINE-ABOVE-COST"><span class="term"><code class="varname">jit_inline_above_cost</code> (<code class="type">floating point</code>)324 <a id="id-1.6.7.10.3.4.12.1.3" class="indexterm"></a>325 </span> <a href="#GUC-JIT-INLINE-ABOVE-COST" class="id_link">#</a></dt><dd><p>326 Sets the query cost above which JIT compilation attempts to inline327 functions and operators. Inlining adds planning time, but can328 improve execution speed. It is not meaningful to set this to less329 than <code class="varname">jit_above_cost</code>.330 Setting this to <code class="literal">-1</code> disables inlining.331 The default is <code class="literal">500000</code>.332 </p></dd><dt id="GUC-JIT-OPTIMIZE-ABOVE-COST"><span class="term"><code class="varname">jit_optimize_above_cost</code> (<code class="type">floating point</code>)333 <a id="id-1.6.7.10.3.4.13.1.3" class="indexterm"></a>334 </span> <a href="#GUC-JIT-OPTIMIZE-ABOVE-COST" class="id_link">#</a></dt><dd><p>335 Sets the query cost above which JIT compilation applies expensive336 optimizations. Such optimization adds planning time, but can improve337 execution speed. It is not meaningful to set this to less338 than <code class="varname">jit_above_cost</code>, and it is unlikely to be339 beneficial to set it to more340 than <code class="varname">jit_inline_above_cost</code>.341 Setting this to <code class="literal">-1</code> disables expensive optimizations.342 The default is <code class="literal">500000</code>.343 </p></dd></dl></div></div><div class="sect2" id="RUNTIME-CONFIG-QUERY-GEQO"><div class="titlepage"><div><div><h3 class="title">20.7.3. Genetic Query Optimizer <a href="#RUNTIME-CONFIG-QUERY-GEQO" class="id_link">#</a></h3></div></div></div><p>344 The genetic query optimizer (GEQO) is an algorithm that does query345 planning using heuristic searching. This reduces planning time for346 complex queries (those joining many relations), at the cost of producing347 plans that are sometimes inferior to those found by the normal348 exhaustive-search algorithm.349 For more information see <a class="xref" href="geqo.html" title="Chapter 62. Genetic Query Optimizer">Chapter 62</a>.350 </p><div class="variablelist"><dl class="variablelist"><dt id="GUC-GEQO"><span class="term"><code class="varname">geqo</code> (<code class="type">boolean</code>)351 <a id="id-1.6.7.10.4.3.1.1.3" class="indexterm"></a>352 <a id="id-1.6.7.10.4.3.1.1.4" class="indexterm"></a>353 <a id="id-1.6.7.10.4.3.1.1.5" class="indexterm"></a>354 </span> <a href="#GUC-GEQO" class="id_link">#</a></dt><dd><p>355 Enables or disables genetic query optimization.356 This is on by default. It is usually best not to turn it off in357 production; the <code class="varname">geqo_threshold</code> variable provides358 more granular control of GEQO.359 </p></dd><dt id="GUC-GEQO-THRESHOLD"><span class="term"><code class="varname">geqo_threshold</code> (<code class="type">integer</code>)360 <a id="id-1.6.7.10.4.3.2.1.3" class="indexterm"></a>361 </span> <a href="#GUC-GEQO-THRESHOLD" class="id_link">#</a></dt><dd><p>362 Use genetic query optimization to plan queries with at least363 this many <code class="literal">FROM</code> items involved. (Note that a364 <code class="literal">FULL OUTER JOIN</code> construct counts as only one <code class="literal">FROM</code>365 item.) The default is 12. For simpler queries it is usually best366 to use the regular, exhaustive-search planner, but for queries with367 many tables the exhaustive search takes too long, often368 longer than the penalty of executing a suboptimal plan. Thus,369 a threshold on the size of the query is a convenient way to manage370 use of GEQO.371 </p></dd><dt id="GUC-GEQO-EFFORT"><span class="term"><code class="varname">geqo_effort</code> (<code class="type">integer</code>)372 <a id="id-1.6.7.10.4.3.3.1.3" class="indexterm"></a>373 </span> <a href="#GUC-GEQO-EFFORT" class="id_link">#</a></dt><dd><p>374 Controls the trade-off between planning time and query plan375 quality in GEQO. This variable must be an integer in the376 range from 1 to 10. The default value is five. Larger values377 increase the time spent doing query planning, but also378 increase the likelihood that an efficient query plan will be379 chosen.380 </p><p>381 <code class="varname">geqo_effort</code> doesn't actually do anything382 directly; it is only used to compute the default values for383 the other variables that influence GEQO behavior (described384 below). If you prefer, you can set the other parameters by385 hand instead.386 </p></dd><dt id="GUC-GEQO-POOL-SIZE"><span class="term"><code class="varname">geqo_pool_size</code> (<code class="type">integer</code>)387 <a id="id-1.6.7.10.4.3.4.1.3" class="indexterm"></a>388 </span> <a href="#GUC-GEQO-POOL-SIZE" class="id_link">#</a></dt><dd><p>389 Controls the pool size used by GEQO, that is the390 number of individuals in the genetic population. It must be391 at least two, and useful values are typically 100 to 1000. If392 it is set to zero (the default setting) then a suitable393 value is chosen based on <code class="varname">geqo_effort</code> and394 the number of tables in the query.395 </p></dd><dt id="GUC-GEQO-GENERATIONS"><span class="term"><code class="varname">geqo_generations</code> (<code class="type">integer</code>)396 <a id="id-1.6.7.10.4.3.5.1.3" class="indexterm"></a>397 </span> <a href="#GUC-GEQO-GENERATIONS" class="id_link">#</a></dt><dd><p>398 Controls the number of generations used by GEQO, that is399 the number of iterations of the algorithm. It must400 be at least one, and useful values are in the same range as401 the pool size. If it is set to zero (the default setting)402 then a suitable value is chosen based on403 <code class="varname">geqo_pool_size</code>.404 </p></dd><dt id="GUC-GEQO-SELECTION-BIAS"><span class="term"><code class="varname">geqo_selection_bias</code> (<code class="type">floating point</code>)405 <a id="id-1.6.7.10.4.3.6.1.3" class="indexterm"></a>406 </span> <a href="#GUC-GEQO-SELECTION-BIAS" class="id_link">#</a></dt><dd><p>407 Controls the selection bias used by GEQO. The selection bias408 is the selective pressure within the population. Values can be409 from 1.50 to 2.00; the latter is the default.410 </p></dd><dt id="GUC-GEQO-SEED"><span class="term"><code class="varname">geqo_seed</code> (<code class="type">floating point</code>)411 <a id="id-1.6.7.10.4.3.7.1.3" class="indexterm"></a>412 </span> <a href="#GUC-GEQO-SEED" class="id_link">#</a></dt><dd><p>413 Controls the initial value of the random number generator used414 by GEQO to select random paths through the join order search space.415 The value can range from zero (the default) to one. Varying the416 value changes the set of join paths explored, and may result in a417 better or worse best path being found.418 </p></dd></dl></div></div><div class="sect2" id="RUNTIME-CONFIG-QUERY-OTHER"><div class="titlepage"><div><div><h3 class="title">20.7.4. Other Planner Options <a href="#RUNTIME-CONFIG-QUERY-OTHER" class="id_link">#</a></h3></div></div></div><div class="variablelist"><dl class="variablelist"><dt id="GUC-DEFAULT-STATISTICS-TARGET"><span class="term"><code class="varname">default_statistics_target</code> (<code class="type">integer</code>)419 <a id="id-1.6.7.10.5.2.1.1.3" class="indexterm"></a>420 </span> <a href="#GUC-DEFAULT-STATISTICS-TARGET" class="id_link">#</a></dt><dd><p>421 Sets the default statistics target for table columns without422 a column-specific target set via <code class="command">ALTER TABLE423 SET STATISTICS</code>. Larger values increase the time needed to424 do <code class="command">ANALYZE</code>, but might improve the quality of the425 planner's estimates. The default is 100. For more information426 on the use of statistics by the <span class="productname">PostgreSQL</span>427 query planner, refer to <a class="xref" href="planner-stats.html" title="14.2. Statistics Used by the Planner">Section 14.2</a>.428 </p></dd><dt id="GUC-CONSTRAINT-EXCLUSION"><span class="term"><code class="varname">constraint_exclusion</code> (<code class="type">enum</code>)429 <a id="id-1.6.7.10.5.2.2.1.3" class="indexterm"></a>430 <a id="id-1.6.7.10.5.2.2.1.4" class="indexterm"></a>431 </span> <a href="#GUC-CONSTRAINT-EXCLUSION" class="id_link">#</a></dt><dd><p>432 Controls the query planner's use of table constraints to433 optimize queries.434 The allowed values of <code class="varname">constraint_exclusion</code> are435 <code class="literal">on</code> (examine constraints for all tables),436 <code class="literal">off</code> (never examine constraints), and437 <code class="literal">partition</code> (examine constraints only for inheritance438 child tables and <code class="literal">UNION ALL</code> subqueries).439 <code class="literal">partition</code> is the default setting.440 It is often used with traditional inheritance trees to improve441 performance.442 </p><p>443 When this parameter allows it for a particular table, the planner444 compares query conditions with the table's <code class="literal">CHECK</code>445 constraints, and omits scanning tables for which the conditions446 contradict the constraints. For example:447 448</p><pre class="programlisting">449CREATE TABLE parent(key integer, ...);450CREATE TABLE child1000(check (key between 1000 and 1999)) INHERITS(parent);451CREATE TABLE child2000(check (key between 2000 and 2999)) INHERITS(parent);452...453SELECT * FROM parent WHERE key = 2400;454</pre><p>455 456 With constraint exclusion enabled, this <code class="command">SELECT</code>457 will not scan <code class="structname">child1000</code> at all, improving performance.458 </p><p>459 Currently, constraint exclusion is enabled by default460 only for cases that are often used to implement table partitioning via461 inheritance trees. Turning it on for all tables imposes extra462 planning overhead that is quite noticeable on simple queries, and most463 often will yield no benefit for simple queries. If you have no464 tables that are partitioned using traditional inheritance, you might465 prefer to turn it off entirely. (Note that the equivalent feature for466 partitioned tables is controlled by a separate parameter,467 <a class="xref" href="runtime-config-query.html#GUC-ENABLE-PARTITION-PRUNING">enable_partition_pruning</a>.)468 </p><p>469 Refer to <a class="xref" href="ddl-partitioning.html#DDL-PARTITIONING-CONSTRAINT-EXCLUSION" title="5.11.5. Partitioning and Constraint Exclusion">Section 5.11.5</a> for470 more information on using constraint exclusion to implement471 partitioning.472 </p></dd><dt id="GUC-CURSOR-TUPLE-FRACTION"><span class="term"><code class="varname">cursor_tuple_fraction</code> (<code class="type">floating point</code>)473 <a id="id-1.6.7.10.5.2.3.1.3" class="indexterm"></a>474 </span> <a href="#GUC-CURSOR-TUPLE-FRACTION" class="id_link">#</a></dt><dd><p>475 Sets the planner's estimate of the fraction of a cursor's rows that476 will be retrieved. The default is 0.1. Smaller values of this477 setting bias the planner towards using <span class="quote">“<span class="quote">fast start</span>”</span> plans478 for cursors, which will retrieve the first few rows quickly while479 perhaps taking a long time to fetch all rows. Larger values480 put more emphasis on the total estimated time. At the maximum481 setting of 1.0, cursors are planned exactly like regular queries,482 considering only the total estimated time and not how soon the483 first rows might be delivered.484 </p></dd><dt id="GUC-FROM-COLLAPSE-LIMIT"><span class="term"><code class="varname">from_collapse_limit</code> (<code class="type">integer</code>)485 <a id="id-1.6.7.10.5.2.4.1.3" class="indexterm"></a>486 </span> <a href="#GUC-FROM-COLLAPSE-LIMIT" class="id_link">#</a></dt><dd><p>487 The planner will merge sub-queries into upper queries if the488 resulting <code class="literal">FROM</code> list would have no more than489 this many items. Smaller values reduce planning time but might490 yield inferior query plans. The default is eight.491 For more information see <a class="xref" href="explicit-joins.html" title="14.3. Controlling the Planner with Explicit JOIN Clauses">Section 14.3</a>.492 </p><p>493 Setting this value to <a class="xref" href="runtime-config-query.html#GUC-GEQO-THRESHOLD">geqo_threshold</a> or more494 may trigger use of the GEQO planner, resulting in non-optimal495 plans. See <a class="xref" href="runtime-config-query.html#RUNTIME-CONFIG-QUERY-GEQO" title="20.7.3. Genetic Query Optimizer">Section 20.7.3</a>.496 </p></dd><dt id="GUC-JIT"><span class="term"><code class="varname">jit</code> (<code class="type">boolean</code>)497 <a id="id-1.6.7.10.5.2.5.1.3" class="indexterm"></a>498 </span> <a href="#GUC-JIT" class="id_link">#</a></dt><dd><p>499 Determines whether <acronym class="acronym">JIT</acronym> compilation may be used by500 <span class="productname">PostgreSQL</span>, if available (see <a class="xref" href="jit.html" title="Chapter 32. Just-in-Time Compilation (JIT)">Chapter 32</a>).501 The default is <code class="literal">on</code>.502 </p></dd><dt id="GUC-JOIN-COLLAPSE-LIMIT"><span class="term"><code class="varname">join_collapse_limit</code> (<code class="type">integer</code>)503 <a id="id-1.6.7.10.5.2.6.1.3" class="indexterm"></a>504 </span> <a href="#GUC-JOIN-COLLAPSE-LIMIT" class="id_link">#</a></dt><dd><p>505 The planner will rewrite explicit <code class="literal">JOIN</code>506 constructs (except <code class="literal">FULL JOIN</code>s) into lists of507 <code class="literal">FROM</code> items whenever a list of no more than this many items508 would result. Smaller values reduce planning time but might509 yield inferior query plans.510 </p><p>511 By default, this variable is set the same as512 <code class="varname">from_collapse_limit</code>, which is appropriate513 for most uses. Setting it to 1 prevents any reordering of514 explicit <code class="literal">JOIN</code>s. Thus, the explicit join order515 specified in the query will be the actual order in which the516 relations are joined. Because the query planner does not always choose517 the optimal join order, advanced users can elect to518 temporarily set this variable to 1, and then specify the join519 order they desire explicitly.520 For more information see <a class="xref" href="explicit-joins.html" title="14.3. Controlling the Planner with Explicit JOIN Clauses">Section 14.3</a>.521 </p><p>522 Setting this value to <a class="xref" href="runtime-config-query.html#GUC-GEQO-THRESHOLD">geqo_threshold</a> or more523 may trigger use of the GEQO planner, resulting in non-optimal524 plans. See <a class="xref" href="runtime-config-query.html#RUNTIME-CONFIG-QUERY-GEQO" title="20.7.3. Genetic Query Optimizer">Section 20.7.3</a>.525 </p></dd><dt id="GUC-PLAN-CACHE-MODE"><span class="term"><code class="varname">plan_cache_mode</code> (<code class="type">enum</code>)526 <a id="id-1.6.7.10.5.2.7.1.3" class="indexterm"></a>527 </span> <a href="#GUC-PLAN-CACHE-MODE" class="id_link">#</a></dt><dd><p>528 Prepared statements (either explicitly prepared or implicitly529 generated, for example by PL/pgSQL) can be executed using custom or530 generic plans. Custom plans are made afresh for each execution531 using its specific set of parameter values, while generic plans do532 not rely on the parameter values and can be re-used across533 executions. Thus, use of a generic plan saves planning time, but if534 the ideal plan depends strongly on the parameter values then a535 generic plan may be inefficient. The choice between these options536 is normally made automatically, but it can be overridden537 with <code class="varname">plan_cache_mode</code>.538 The allowed values are <code class="literal">auto</code> (the default),539 <code class="literal">force_custom_plan</code> and540 <code class="literal">force_generic_plan</code>.541 This setting is considered when a cached plan is to be executed,542 not when it is prepared.543 For more information see <a class="xref" href="sql-prepare.html" title="PREPARE"><span class="refentrytitle">PREPARE</span></a>.544 </p></dd><dt id="GUC-RECURSIVE-WORKTABLE-FACTOR"><span class="term"><code class="varname">recursive_worktable_factor</code> (<code class="type">floating point</code>)545 <a id="id-1.6.7.10.5.2.8.1.3" class="indexterm"></a>546 </span> <a href="#GUC-RECURSIVE-WORKTABLE-FACTOR" class="id_link">#</a></dt><dd><p>547 Sets the planner's estimate of the average size of the working548 table of a <a class="link" href="queries-with.html#QUERIES-WITH-RECURSIVE" title="7.8.2. Recursive Queries">recursive549 query</a>, as a multiple of the estimated size of the initial550 non-recursive term of the query. This helps the planner choose551 the most appropriate method for joining the working table to the552 query's other tables.553 The default value is <code class="literal">10.0</code>. A smaller value554 such as <code class="literal">1.0</code> can be helpful when the recursion555 has low <span class="quote">“<span class="quote">fan-out</span>”</span> from one step to the next, as for556 example in shortest-path queries. Graph analytics queries may557 benefit from larger-than-default values.558 </p></dd></dl></div></div></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="runtime-config-replication.html" title="20.6. Replication">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="runtime-config.html" title="Chapter 20. Server Configuration">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="runtime-config-logging.html" title="20.8. Error Reporting and Logging">Next</a></td></tr><tr><td width="40%" align="left" valign="top">20.6. Replication </td><td width="20%" align="center"><a accesskey="h" href="index.html" title="PostgreSQL 16.3 Documentation">Home</a></td><td width="40%" align="right" valign="top"> 20.8. Error Reporting and Logging</td></tr></table></div></body></html>