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>11.2. Index Types</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="indexes-intro.html" title="11.1. Introduction" /><link rel="next" href="indexes-multicolumn.html" title="11.3. Multicolumn Indexes" /></head><body id="docContent" class="container-fluid col-10"><div class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">11.2. Index Types</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="indexes-intro.html" title="11.1. Introduction">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="indexes.html" title="Chapter 11. Indexes">Up</a></td><th width="60%" align="center">Chapter 11. Indexes</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="indexes-multicolumn.html" title="11.3. Multicolumn Indexes">Next</a></td></tr></table><hr /></div><div class="sect1" id="INDEXES-TYPES"><div class="titlepage"><div><div><h2 class="title" style="clear: both">11.2. Index Types <a href="#INDEXES-TYPES" class="id_link">#</a></h2></div></div></div><div class="toc"><dl class="toc"><dt><span class="sect2"><a href="indexes-types.html#INDEXES-TYPES-BTREE">11.2.1. B-Tree</a></span></dt><dt><span class="sect2"><a href="indexes-types.html#INDEXES-TYPES-HASH">11.2.2. Hash</a></span></dt><dt><span class="sect2"><a href="indexes-types.html#INDEXES-TYPE-GIST">11.2.3. GiST</a></span></dt><dt><span class="sect2"><a href="indexes-types.html#INDEXES-TYPE-SPGIST">11.2.4. SP-GiST</a></span></dt><dt><span class="sect2"><a href="indexes-types.html#INDEXES-TYPES-GIN">11.2.5. GIN</a></span></dt><dt><span class="sect2"><a href="indexes-types.html#INDEXES-TYPES-BRIN">11.2.6. BRIN</a></span></dt></dl></div><p>
3 <span class="productname">PostgreSQL</span> provides several index types:
4 B-tree, Hash, GiST, SP-GiST, GIN, BRIN, and the extension <a class="link" href="bloom.html" title="F.6. bloom — bloom filter index access method">bloom</a>.
5 Each index type uses a different
6 algorithm that is best suited to different types of indexable clauses.
7 By default, the <a class="link" href="sql-createindex.html" title="CREATE INDEX"><code class="command">CREATE
8 INDEX</code></a> command creates
9 B-tree indexes, which fit the most common situations.
10 The other index types are selected by writing the keyword
11 <code class="literal">USING</code> followed by the index type name.
12 For example, to create a Hash index:
13 </p><pre class="programlisting">
14 CREATE INDEX <em class="replaceable"><code>name</code></em> ON <em class="replaceable"><code>table</code></em> USING HASH (<em class="replaceable"><code>column</code></em>);
16 </p><div class="sect2" id="INDEXES-TYPES-BTREE"><div class="titlepage"><div><div><h3 class="title">11.2.1. B-Tree <a href="#INDEXES-TYPES-BTREE" class="id_link">#</a></h3></div></div></div><a id="id-1.5.10.5.3.2" class="indexterm"></a><a id="id-1.5.10.5.3.3" class="indexterm"></a><p>
17 B-trees can handle equality and range queries on data that can be sorted
19 In particular, the <span class="productname">PostgreSQL</span> query planner
20 will consider using a B-tree index whenever an indexed column is
21 involved in a comparison using one of these operators:
23 </p><pre class="synopsis">
24 < <= = >= >
27 Constructs equivalent to combinations of these operators, such as
28 <code class="literal">BETWEEN</code> and <code class="literal">IN</code>, can also be implemented with
29 a B-tree index search. Also, an <code class="literal">IS NULL</code> or <code class="literal">IS NOT
30 NULL</code> condition on an index column can be used with a B-tree index.
32 The optimizer can also use a B-tree index for queries involving the
33 pattern matching operators <code class="literal">LIKE</code> and <code class="literal">~</code>
34 <span class="emphasis"><em>if</em></span> the pattern is a constant and is anchored to
35 the beginning of the string — for example, <code class="literal">col LIKE
36 'foo%'</code> or <code class="literal">col ~ '^foo'</code>, but not
37 <code class="literal">col LIKE '%bar'</code>. However, if your database does not
38 use the C locale you will need to create the index with a special
39 operator class to support indexing of pattern-matching queries; see
40 <a class="xref" href="indexes-opclass.html" title="11.10. Operator Classes and Operator Families">Section 11.10</a> below. It is also possible to use
41 B-tree indexes for <code class="literal">ILIKE</code> and
42 <code class="literal">~*</code>, but only if the pattern starts with
43 non-alphabetic characters, i.e., characters that are not affected by
44 upper/lower case conversion.
46 B-tree indexes can also be used to retrieve data in sorted order.
47 This is not always faster than a simple scan and sort, but it is
49 </p></div><div class="sect2" id="INDEXES-TYPES-HASH"><div class="titlepage"><div><div><h3 class="title">11.2.2. Hash <a href="#INDEXES-TYPES-HASH" class="id_link">#</a></h3></div></div></div><a id="id-1.5.10.5.4.2" class="indexterm"></a><a id="id-1.5.10.5.4.3" class="indexterm"></a><p>
50 Hash indexes store a 32-bit hash code derived from the
51 value of the indexed column. Hence,
52 such indexes can only handle simple equality comparisons.
53 The query planner will consider using a hash index whenever an
54 indexed column is involved in a comparison using the
57 </p><pre class="synopsis">
60 </p></div><div class="sect2" id="INDEXES-TYPE-GIST"><div class="titlepage"><div><div><h3 class="title">11.2.3. GiST <a href="#INDEXES-TYPE-GIST" class="id_link">#</a></h3></div></div></div><a id="id-1.5.10.5.5.2" class="indexterm"></a><a id="id-1.5.10.5.5.3" class="indexterm"></a><p>
61 GiST indexes are not a single kind of index, but rather an infrastructure
62 within which many different indexing strategies can be implemented.
63 Accordingly, the particular operators with which a GiST index can be
64 used vary depending on the indexing strategy (the <em class="firstterm">operator
65 class</em>). As an example, the standard distribution of
66 <span class="productname">PostgreSQL</span> includes GiST operator classes
67 for several two-dimensional geometric data types, which support indexed
68 queries using these operators:
70 </p><pre class="synopsis">
71 << &< &> >> <<| &<| |&> |>> @> <@ ~= &&
74 (See <a class="xref" href="functions-geometry.html" title="9.11. Geometric Functions and Operators">Section 9.11</a> for the meaning of
76 The GiST operator classes included in the standard distribution are
77 documented in <a class="xref" href="gist.html#GIST-BUILTIN-OPCLASSES-TABLE" title="Table 65.1. Built-in GiST Operator Classes">Table 65.1</a>.
78 Many other GiST operator
79 classes are available in the <code class="literal">contrib</code> collection or as separate
80 projects. For more information see <a class="xref" href="gist.html" title="65.2. GiST Indexes">Section 65.2</a>.
82 GiST indexes are also capable of optimizing <span class="quote">“<span class="quote">nearest-neighbor</span>”</span>
84 </p><pre class="programlisting">
85 SELECT * FROM places ORDER BY location <-> point '(101,456)' LIMIT 10;
88 which finds the ten places closest to a given target point. The ability
89 to do this is again dependent on the particular operator class being used.
90 In <a class="xref" href="gist.html#GIST-BUILTIN-OPCLASSES-TABLE" title="Table 65.1. Built-in GiST Operator Classes">Table 65.1</a>, operators that can be
91 used in this way are listed in the column <span class="quote">“<span class="quote">Ordering Operators</span>”</span>.
92 </p></div><div class="sect2" id="INDEXES-TYPE-SPGIST"><div class="titlepage"><div><div><h3 class="title">11.2.4. SP-GiST <a href="#INDEXES-TYPE-SPGIST" class="id_link">#</a></h3></div></div></div><a id="id-1.5.10.5.6.2" class="indexterm"></a><a id="id-1.5.10.5.6.3" class="indexterm"></a><p>
93 SP-GiST indexes, like GiST indexes, offer an infrastructure that supports
94 various kinds of searches. SP-GiST permits implementation of a wide range
95 of different non-balanced disk-based data structures, such as quadtrees,
96 k-d trees, and radix trees (tries). As an example, the standard distribution of
97 <span class="productname">PostgreSQL</span> includes SP-GiST operator classes
98 for two-dimensional points, which support indexed
99 queries using these operators:
101 </p><pre class="synopsis">
102 << >> ~= <@ <<| |>>
105 (See <a class="xref" href="functions-geometry.html" title="9.11. Geometric Functions and Operators">Section 9.11</a> for the meaning of
107 The SP-GiST operator classes included in the standard distribution are
108 documented in <a class="xref" href="spgist.html#SPGIST-BUILTIN-OPCLASSES-TABLE" title="Table 65.2. Built-in SP-GiST Operator Classes">Table 65.2</a>.
109 For more information see <a class="xref" href="spgist.html" title="65.3. SP-GiST Indexes">Section 65.3</a>.
111 Like GiST, SP-GiST supports <span class="quote">“<span class="quote">nearest-neighbor</span>”</span> searches.
112 For SP-GiST operator classes that support distance ordering, the
113 corresponding operator is listed in the <span class="quote">“<span class="quote">Ordering Operators</span>”</span>
114 column in <a class="xref" href="spgist.html#SPGIST-BUILTIN-OPCLASSES-TABLE" title="Table 65.2. Built-in SP-GiST Operator Classes">Table 65.2</a>.
115 </p></div><div class="sect2" id="INDEXES-TYPES-GIN"><div class="titlepage"><div><div><h3 class="title">11.2.5. GIN <a href="#INDEXES-TYPES-GIN" class="id_link">#</a></h3></div></div></div><a id="id-1.5.10.5.7.2" class="indexterm"></a><a id="id-1.5.10.5.7.3" class="indexterm"></a><p>
116 GIN indexes are <span class="quote">“<span class="quote">inverted indexes</span>”</span> which are appropriate for
117 data values that contain multiple component values, such as arrays. An
118 inverted index contains a separate entry for each component value, and
119 can efficiently handle queries that test for the presence of specific
122 Like GiST and SP-GiST, GIN can support
123 many different user-defined indexing strategies, and the particular
124 operators with which a GIN index can be used vary depending on the
126 As an example, the standard distribution of
127 <span class="productname">PostgreSQL</span> includes a GIN operator class
128 for arrays, which supports indexed queries using these operators:
130 </p><pre class="synopsis">
131 <@ @> = &&
134 (See <a class="xref" href="functions-array.html" title="9.19. Array Functions and Operators">Section 9.19</a> for the meaning of
136 The GIN operator classes included in the standard distribution are
137 documented in <a class="xref" href="gin.html#GIN-BUILTIN-OPCLASSES-TABLE" title="Table 65.3. Built-in GIN Operator Classes">Table 65.3</a>.
138 Many other GIN operator
139 classes are available in the <code class="literal">contrib</code> collection or as separate
140 projects. For more information see <a class="xref" href="gin.html" title="65.4. GIN Indexes">Section 65.4</a>.
141 </p></div><div class="sect2" id="INDEXES-TYPES-BRIN"><div class="titlepage"><div><div><h3 class="title">11.2.6. BRIN <a href="#INDEXES-TYPES-BRIN" class="id_link">#</a></h3></div></div></div><a id="id-1.5.10.5.8.2" class="indexterm"></a><a id="id-1.5.10.5.8.3" class="indexterm"></a><p>
142 BRIN indexes (a shorthand for Block Range INdexes) store summaries about
143 the values stored in consecutive physical block ranges of a table.
144 Thus, they are most effective for columns whose values are well-correlated
145 with the physical order of the table rows.
146 Like GiST, SP-GiST and GIN,
147 BRIN can support many different indexing strategies,
148 and the particular operators with which a BRIN index can be used
149 vary depending on the indexing strategy.
150 For data types that have a linear sort order, the indexed data
151 corresponds to the minimum and maximum values of the
152 values in the column for each block range. This supports indexed queries
153 using these operators:
155 </p><pre class="synopsis">
156 < <= = >= >
159 The BRIN operator classes included in the standard distribution are
160 documented in <a class="xref" href="brin.html#BRIN-BUILTIN-OPCLASSES-TABLE" title="Table 65.4. Built-in BRIN Operator Classes">Table 65.4</a>.
161 For more information see <a class="xref" href="brin.html" title="65.5. BRIN Indexes">Section 65.5</a>.
162 </p></div></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="indexes-intro.html" title="11.1. Introduction">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="indexes.html" title="Chapter 11. Indexes">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="indexes-multicolumn.html" title="11.3. Multicolumn Indexes">Next</a></td></tr><tr><td width="40%" align="left" valign="top">11.1. Introduction </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"> 11.3. Multicolumn Indexes</td></tr></table></div></body></html>