2 $PostgreSQL: pgsql/doc/src/sgml/ref/create_index.sgml,v 1.50 2004/11/05 19:15:51 tgl Exp $
3 PostgreSQL documentation
6 <refentry id="SQL-CREATEINDEX">
8 <refentrytitle id="sql-createindex-title">CREATE INDEX</refentrytitle>
9 <refmiscinfo>SQL - Language Statements</refmiscinfo>
13 <refname>CREATE INDEX</refname>
14 <refpurpose>define a new index</refpurpose>
17 <indexterm zone="sql-createindex">
18 <primary>CREATE INDEX</primary>
23 CREATE [ UNIQUE ] INDEX <replaceable class="parameter">name</replaceable> ON <replaceable class="parameter">table</replaceable> [ USING <replaceable class="parameter">method</replaceable> ]
24 ( { <replaceable class="parameter">column</replaceable> | ( <replaceable class="parameter">expression</replaceable> ) } [ <replaceable class="parameter">opclass</replaceable> ] [, ...] )
25 [ TABLESPACE <replaceable class="parameter">tablespace</replaceable> ]
26 [ WHERE <replaceable class="parameter">predicate</replaceable> ]
31 <title>Description</title>
34 <command>CREATE INDEX</command> constructs an index <replaceable
35 class="parameter">index_name</replaceable> on the specified table.
36 Indexes are primarily used to enhance database performance (though
37 inappropriate use will result in slower performance).
41 The key field(s) for the index are specified as column names,
42 or alternatively as expressions written in parentheses.
43 Multiple fields can be specified if the index method supports
48 An index field can be an expression computed from the values of
49 one or more columns of the table row. This feature can be used
50 to obtain fast access to data based on some transformation of
51 the basic data. For example, an index computed on
52 <literal>upper(col)</> would allow the clause
53 <literal>WHERE upper(col) = 'JIM'</> to use an index.
57 <productname>PostgreSQL</productname> provides the index methods
58 B-tree, R-tree, hash, and GiST. The B-tree index method is an
59 implementation of Lehman-Yao high-concurrency B-trees. The R-tree
60 index method implements standard R-trees using Guttman's quadratic
61 split algorithm. The hash index method is an implementation of
62 Litwin's linear hashing. Users can also define their own index
63 methods, but that is fairly complicated.
67 When the <literal>WHERE</literal> clause is present, a
68 <firstterm>partial index</firstterm> is created.
69 A partial index is an index that contains entries for only a portion of
70 a table, usually a portion that is more useful for indexing than the
71 rest of the table. For example, if you have a table that contains both
72 billed and unbilled orders where the unbilled orders take up a small
73 fraction of the total table and yet that is an often used section, you
74 can improve performance by creating an index on just that portion.
75 Another possible application is to use <literal>WHERE</literal> with
76 <literal>UNIQUE</literal> to enforce uniqueness over a subset of a
81 Indexes are not used for <literal>IS NULL</> clauses by default.
82 The best way to use indexes in such cases is to create a partial index
83 using an <literal>IS NULL</> comparison.
87 The expression used in the <literal>WHERE</literal> clause may refer
88 only to columns of the underlying table, but it can use all columns,
89 not just the ones being indexed. Presently, subqueries and
90 aggregate expressions are also forbidden in <literal>WHERE</literal>.
91 The same restrictions apply to index fields that are expressions.
95 All functions and operators used in an index definition must be
96 <quote>immutable</>, that is, their results must depend only on
97 their arguments and never on any outside influence (such as
98 the contents of another table or the current time). This restriction
99 ensures that the behavior of the index is well-defined. To use a
100 user-defined function in an index expression or <literal>WHERE</literal>
101 clause, remember to mark the function immutable when you create it.
106 <title>Parameters</title>
110 <term><literal>UNIQUE</literal></term>
113 Causes the system to check for
114 duplicate values in the table when the index is created (if data
115 already exist) and each time data is added. Attempts to
116 insert or update data which would result in duplicate entries
117 will generate an error.
123 <term><replaceable class="parameter">name</replaceable></term>
126 The name of the index to be created. No schema name can be included
127 here; the index is always created in the same schema as its parent
134 <term><replaceable class="parameter">table</replaceable></term>
137 The name (possibly schema-qualified) of the table to be indexed.
143 <term><replaceable class="parameter">method</replaceable></term>
146 The name of the method to be used for the index. Choices are
147 <literal>btree</literal>, <literal>hash</literal>,
148 <literal>rtree</literal>, and <literal>gist</literal>. The
149 default method is <literal>btree</literal>.
155 <term><replaceable class="parameter">column</replaceable></term>
158 The name of a column of the table.
164 <term><replaceable class="parameter">expression</replaceable></term>
167 An expression based on one or more columns of the table. The
168 expression usually must be written with surrounding parentheses,
169 as shown in the syntax. However, the parentheses may be omitted
170 if the expression has the form of a function call.
176 <term><replaceable class="parameter">opclass</replaceable></term>
179 The name of an operator class. See below for details.
185 <term><replaceable class="parameter">tablespace</replaceable></term>
188 The tablespace in which to create the index. If not specified,
189 <xref linkend="guc-default-tablespace"> is used, or the database's
190 default tablespace if <varname>default_tablespace</> is an empty
197 <term><replaceable class="parameter">predicate</replaceable></term>
200 The constraint expression for a partial index.
212 See <xref linkend="indexes"> for information about when indexes can
213 be used, when they are not used, and in which particular situations
218 Currently, only the B-tree and GiST index methods support
219 multicolumn indexes. Up to 32 fields may be specified by default.
220 (This limit can be altered when building
221 <productname>PostgreSQL</productname>.) Only B-tree currently
222 supports unique indexes.
226 An <firstterm>operator class</firstterm> can be specified for each
227 column of an index. The operator class identifies the operators to be
228 used by the index for that column. For example, a B-tree index on
229 four-byte integers would use the <literal>int4_ops</literal> class;
230 this operator class includes comparison functions for four-byte
231 integers. In practice the default operator class for the column's data
232 type is usually sufficient. The main point of having operator classes
233 is that for some data types, there could be more than one meaningful
234 ordering. For example, we might want to sort a complex-number data
235 type either by absolute value or by real part. We could do this by
236 defining two operator classes for the data type and then selecting
237 the proper class when making an index. More information about
238 operator classes is in <xref linkend="indexes-opclass"> and in <xref
243 Use <xref linkend="sql-dropindex" endterm="sql-dropindex-title">
249 <title>Examples</title>
252 To create a B-tree index on the column <literal>title</literal> in
253 the table <literal>films</literal>:
255 CREATE UNIQUE INDEX title_idx ON films (title);
260 To create an index on the column <literal>code</> in the table
261 <literal>films</> and have the index reside in the tablespace
262 <literal>indexspace</>:
264 CREATE INDEX code_idx ON films(code) TABLESPACE indexspace;
270 Is this example correct?
273 To create a R-tree index on a point attribute so that we
274 can efficiently use box operators on the result of the
278 CREATE INDEX pointloc
279 ON points USING RTREE (point2box(location) box_ops);
281 WHERE point2box(points.pointloc) = boxes.box;
288 <title>Compatibility</title>
291 <command>CREATE INDEX</command> is a
292 <productname>PostgreSQL</productname> language extension. There
293 are no provisions for indexes in the SQL standard.
298 <!-- Keep this comment at the end of the file
303 sgml-minimize-attributes:nil
304 sgml-always-quote-attributes:t
307 sgml-parent-document:nil
308 sgml-default-dtd-file:"../reference.ced"
309 sgml-exposed-tags:nil
310 sgml-local-catalogs:"/usr/lib/sgml/catalog"
311 sgml-local-ecat-files:nil