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>CLUSTER</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="sql-close.html" title="CLOSE" /><link rel="next" href="sql-comment.html" title="COMMENT" /></head><body id="docContent" class="container-fluid col-10"><div class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">CLUSTER</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="sql-close.html" title="CLOSE">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="sql-commands.html" title="SQL Commands">Up</a></td><th width="60%" align="center">SQL Commands</th><td width="10%" align="right"><a accesskey="h" href="index.html" title="PostgreSQL 18.0 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="sql-comment.html" title="COMMENT">Next</a></td></tr></table><hr /></div><div class="refentry" id="SQL-CLUSTER"><div class="titlepage"></div><a id="id-1.9.3.51.1" class="indexterm"></a><div class="refnamediv"><h2><span class="refentrytitle">CLUSTER</span></h2><p>CLUSTER — cluster a table according to an index</p></div><div class="refsynopsisdiv"><h2>Synopsis</h2><pre class="synopsis">
3 CLUSTER [ ( <em class="replaceable"><code>option</code></em> [, ...] ) ] [ <em class="replaceable"><code>table_name</code></em> [ USING <em class="replaceable"><code>index_name</code></em> ] ]
5 <span class="phrase">where <em class="replaceable"><code>option</code></em> can be one of:</span>
7 VERBOSE [ <em class="replaceable"><code>boolean</code></em> ]
8 </pre></div><div class="refsect1" id="id-1.9.3.51.5"><h2>Description</h2><p>
9 <code class="command">CLUSTER</code> instructs <span class="productname">PostgreSQL</span>
10 to cluster the table specified
11 by <em class="replaceable"><code>table_name</code></em>
12 based on the index specified by
13 <em class="replaceable"><code>index_name</code></em>. The index must
14 already have been defined on
15 <em class="replaceable"><code>table_name</code></em>.
17 When a table is clustered, it is physically reordered
18 based on the index information. Clustering is a one-time operation:
19 when the table is subsequently updated, the changes are
20 not clustered. That is, no attempt is made to store new or
21 updated rows according to their index order. (If one wishes, one can
22 periodically recluster by issuing the command again. Also, setting
23 the table's <code class="literal">fillfactor</code> storage parameter to less than
24 100% can aid in preserving cluster ordering during updates, since updated
25 rows are kept on the same page if enough space is available there.)
27 When a table is clustered, <span class="productname">PostgreSQL</span>
28 remembers which index it was clustered by. The form
29 <code class="command">CLUSTER <em class="replaceable"><code>table_name</code></em></code>
30 reclusters the table using the same index as before. You can also
31 use the <code class="literal">CLUSTER</code> or <code class="literal">SET WITHOUT CLUSTER</code>
32 forms of <a class="link" href="sql-altertable.html" title="ALTER TABLE"><code class="command">ALTER TABLE</code></a> to set the index to be used for
33 future cluster operations, or to clear any previous setting.
35 <code class="command">CLUSTER</code> without a
36 <em class="replaceable"><code>table_name</code></em> reclusters all the
37 previously-clustered tables in the current database that the calling user
38 has privileges for. This form of <code class="command">CLUSTER</code> cannot be
39 executed inside a transaction block.
41 When a table is being clustered, an <code class="literal">ACCESS
42 EXCLUSIVE</code> lock is acquired on it. This prevents any other
43 database operations (both reads and writes) from operating on the
44 table until the <code class="command">CLUSTER</code> is finished.
45 </p></div><div class="refsect1" id="id-1.9.3.51.6"><h2>Parameters</h2><div class="variablelist"><dl class="variablelist"><dt><span class="term"><em class="replaceable"><code>table_name</code></em></span></dt><dd><p>
46 The name (possibly schema-qualified) of a table.
47 </p></dd><dt><span class="term"><em class="replaceable"><code>index_name</code></em></span></dt><dd><p>
49 </p></dd><dt><span class="term"><code class="literal">VERBOSE</code></span></dt><dd><p>
50 Prints a progress report as each table is clustered
51 at <code class="literal">INFO</code> level.
52 </p></dd><dt><span class="term"><em class="replaceable"><code>boolean</code></em></span></dt><dd><p>
53 Specifies whether the selected option should be turned on or off.
54 You can write <code class="literal">TRUE</code>, <code class="literal">ON</code>, or
55 <code class="literal">1</code> to enable the option, and <code class="literal">FALSE</code>,
56 <code class="literal">OFF</code>, or <code class="literal">0</code> to disable it. The
57 <em class="replaceable"><code>boolean</code></em> value can also
58 be omitted, in which case <code class="literal">TRUE</code> is assumed.
59 </p></dd></dl></div></div><div class="refsect1" id="id-1.9.3.51.7"><h2>Notes</h2><p>
60 To cluster a table, one must have the <code class="literal">MAINTAIN</code> privilege
63 In cases where you are accessing single rows randomly
64 within a table, the actual order of the data in the
65 table is unimportant. However, if you tend to access some
66 data more than others, and there is an index that groups
67 them together, you will benefit from using <code class="command">CLUSTER</code>.
68 If you are requesting a range of indexed values from a table, or a
69 single indexed value that has multiple rows that match,
70 <code class="command">CLUSTER</code> will help because once the index identifies the
71 table page for the first row that matches, all other rows
72 that match are probably already on the same table page,
73 and so you save disk accesses and speed up the query.
75 <code class="command">CLUSTER</code> can re-sort the table using either an index scan
76 on the specified index, or (if the index is a b-tree) a sequential
77 scan followed by sorting. It will attempt to choose the method that
78 will be faster, based on planner cost parameters and available statistical
81 While <code class="command">CLUSTER</code> is running, the <a class="xref" href="runtime-config-client.html#GUC-SEARCH-PATH">search_path</a> is temporarily changed to <code class="literal">pg_catalog,
84 When an index scan is used, a temporary copy of the table is created that
85 contains the table data in the index order. Temporary copies of each
86 index on the table are created as well. Therefore, you need free space on
87 disk at least equal to the sum of the table size and the index sizes.
89 When a sequential scan and sort is used, a temporary sort file is
90 also created, so that the peak temporary space requirement is as much
91 as double the table size, plus the index sizes. This method is often
92 faster than the index scan method, but if the disk space requirement is
93 intolerable, you can disable this choice by temporarily setting <a class="xref" href="runtime-config-query.html#GUC-ENABLE-SORT">enable_sort</a> to <code class="literal">off</code>.
95 It is advisable to set <a class="xref" href="runtime-config-resource.html#GUC-MAINTENANCE-WORK-MEM">maintenance_work_mem</a> to
96 a reasonably large value (but not more than the amount of RAM you can
97 dedicate to the <code class="command">CLUSTER</code> operation) before clustering.
99 Because the planner records statistics about the ordering of
100 tables, it is advisable to run <a class="link" href="sql-analyze.html" title="ANALYZE"><code class="command">ANALYZE</code></a>
101 on the newly clustered table.
102 Otherwise, the planner might make poor choices of query plans.
104 Because <code class="command">CLUSTER</code> remembers which indexes are clustered,
105 one can cluster the tables one wants clustered manually the first time,
106 then set up a periodic maintenance script that executes
107 <code class="command">CLUSTER</code> without any parameters, so that the desired tables
108 are periodically reclustered.
110 Each backend running <code class="command">CLUSTER</code> will report its progress
111 in the <code class="structname">pg_stat_progress_cluster</code> view. See
112 <a class="xref" href="progress-reporting.html#CLUSTER-PROGRESS-REPORTING" title="27.4.2. CLUSTER Progress Reporting">Section 27.4.2</a> for details.
114 Clustering a partitioned table clusters each of its partitions using the
115 partition of the specified partitioned index. When clustering a partitioned
116 table, the index may not be omitted. <code class="command">CLUSTER</code> on a
117 partitioned table cannot be executed inside a transaction block.
118 </p></div><div class="refsect1" id="id-1.9.3.51.8"><h2>Examples</h2><p>
119 Cluster the table <code class="literal">employees</code> on the basis of
120 its index <code class="literal">employees_ind</code>:
121 </p><pre class="programlisting">
122 CLUSTER employees USING employees_ind;
125 Cluster the <code class="literal">employees</code> table using the same
126 index that was used before:
127 </p><pre class="programlisting">
131 Cluster all tables in the database that have previously been clustered:
132 </p><pre class="programlisting">
134 </pre></div><div class="refsect1" id="id-1.9.3.51.9"><h2>Compatibility</h2><p>
135 There is no <code class="command">CLUSTER</code> statement in the SQL standard.
137 The following syntax was used before <span class="productname">PostgreSQL</span>
138 17 and is still supported:
139 </p><pre class="synopsis">
140 CLUSTER [ VERBOSE ] [ <em class="replaceable"><code>table_name</code></em> [ USING <em class="replaceable"><code>index_name</code></em> ] ]
143 The following syntax was used before <span class="productname">PostgreSQL</span>
144 8.3 and is still supported:
145 </p><pre class="synopsis">
146 CLUSTER <em class="replaceable"><code>index_name</code></em> ON <em class="replaceable"><code>table_name</code></em>
148 </p></div><div class="refsect1" id="id-1.9.3.51.10"><h2>See Also</h2><span class="simplelist"><a class="xref" href="app-clusterdb.html" title="clusterdb"><span class="refentrytitle"><span class="application">clusterdb</span></span></a>, <a class="xref" href="progress-reporting.html#CLUSTER-PROGRESS-REPORTING" title="27.4.2. CLUSTER Progress Reporting">Section 27.4.2</a></span></div></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="sql-close.html" title="CLOSE">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="sql-commands.html" title="SQL Commands">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="sql-comment.html" title="COMMENT">Next</a></td></tr><tr><td width="40%" align="left" valign="top">CLOSE </td><td width="20%" align="center"><a accesskey="h" href="index.html" title="PostgreSQL 18.0 Documentation">Home</a></td><td width="40%" align="right" valign="top"> COMMENT</td></tr></table></div></body></html>