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>Chapter 63. Index Access Method Interface Definition</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="tableam.html" title="Chapter 62. Table Access Method Interface Definition" /><link rel="next" href="index-api.html" title="63.1. Basic API Structure for 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">Chapter 63. Index Access Method Interface Definition</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="tableam.html" title="Chapter 62. Table Access Method Interface Definition">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="internals.html" title="Part VII. Internals">Up</a></td><th width="60%" align="center">Part VII. Internals</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="index-api.html" title="63.1. Basic API Structure for Indexes">Next</a></td></tr></table><hr /></div><div class="chapter" id="INDEXAM"><div class="titlepage"><div><div><h2 class="title">Chapter 63. Index Access Method Interface Definition</h2></div></div></div><div class="toc"><p><strong>Table of Contents</strong></p><dl class="toc"><dt><span class="sect1"><a href="index-api.html">63.1. Basic API Structure for Indexes</a></span></dt><dt><span class="sect1"><a href="index-functions.html">63.2. Index Access Method Functions</a></span></dt><dt><span class="sect1"><a href="index-scanning.html">63.3. Index Scanning</a></span></dt><dt><span class="sect1"><a href="index-locking.html">63.4. Index Locking Considerations</a></span></dt><dt><span class="sect1"><a href="index-unique-checks.html">63.5. Index Uniqueness Checks</a></span></dt><dt><span class="sect1"><a href="index-cost-estimation.html">63.6. Index Cost Estimation Functions</a></span></dt></dl></div><a id="id-1.10.15.2" class="indexterm"></a><a id="id-1.10.15.3" class="indexterm"></a><p>
3 This chapter defines the interface between the core
4 <span class="productname">PostgreSQL</span> system and <em class="firstterm">index access
5 methods</em>, which manage individual index types. The core system
6 knows nothing about indexes beyond what is specified here, so it is
7 possible to develop entirely new index types by writing add-on code.
9 All indexes in <span class="productname">PostgreSQL</span> are what are known
10 technically as <em class="firstterm">secondary indexes</em>; that is, the index is
11 physically separate from the table file that it describes. Each index
12 is stored as its own physical <em class="firstterm">relation</em> and so is described
13 by an entry in the <code class="structname">pg_class</code> catalog. The contents of an
14 index are entirely under the control of its index access method. In
15 practice, all index access methods divide indexes into standard-size
16 pages so that they can use the regular storage manager and buffer manager
17 to access the index contents. (All the existing index access methods
18 furthermore use the standard page layout described in <a class="xref" href="storage-page-layout.html" title="66.6. Database Page Layout">Section 66.6</a>, and most use the same format for index
19 tuple headers; but these decisions are not forced on an access method.)
21 An index is effectively a mapping from some data key values to
22 <em class="firstterm">tuple identifiers</em>, or <acronym class="acronym">TIDs</acronym>, of row versions
23 (tuples) in the index's parent table. A TID consists of a
24 block number and an item number within that block (see <a class="xref" href="storage-page-layout.html" title="66.6. Database Page Layout">Section 66.6</a>). This is sufficient
25 information to fetch a particular row version from the table.
26 Indexes are not directly aware that under MVCC, there might be multiple
27 extant versions of the same logical row; to an index, each tuple is
28 an independent object that needs its own index entry. Thus, an
29 update of a row always creates all-new index entries for the row, even if
30 the key values did not change. (<a class="link" href="storage-hot.html" title="66.7. Heap-Only Tuples (HOT)">HOT
31 tuples</a> are an exception to this
32 statement; but indexes do not deal with those, either.) Index entries for
33 dead tuples are reclaimed (by vacuuming) when the dead tuples themselves
35 </p></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="tableam.html" title="Chapter 62. Table Access Method Interface Definition">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="internals.html" title="Part VII. Internals">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="index-api.html" title="63.1. Basic API Structure for Indexes">Next</a></td></tr><tr><td width="40%" align="left" valign="top">Chapter 62. Table Access Method Interface Definition </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"> 63.1. Basic API Structure for Indexes</td></tr></table></div></body></html>