create_index.sgml 13.8 KB
Newer Older
1
<!--
2
$Header: /cvsroot/pgsql/doc/src/sgml/ref/create_index.sgml,v 1.37 2002/09/21 18:32:54 petere Exp $
3
PostgreSQL documentation
4 5
-->

6 7
<refentry id="SQL-CREATEINDEX">
 <refmeta>
8
  <refentrytitle id="sql-createindex-title">CREATE INDEX</refentrytitle>
9 10 11 12
  <refmiscinfo>SQL - Language Statements</refmiscinfo>
 </refmeta>
 <refnamediv>
  <refname>
13
   CREATE INDEX
14 15
  </refname>
  <refpurpose>
16
   define a new index
17
  </refpurpose>
18
 </refnamediv>
19 20
 <refsynopsisdiv>
  <refsynopsisdivinfo>
21
   <date>2001-07-15</date>
22 23
  </refsynopsisdivinfo>
  <synopsis>
24
CREATE [ UNIQUE ] INDEX <replaceable class="parameter">index_name</replaceable> ON <replaceable class="parameter">table</replaceable>
25 26
    [ USING <replaceable class="parameter">acc_method</replaceable> ] ( <replaceable class="parameter">column</replaceable> [ <replaceable class="parameter">ops_name</replaceable> ] [, ...] )
    [ WHERE <replaceable class="parameter">predicate</replaceable> ]
27
CREATE [ UNIQUE ] INDEX <replaceable class="parameter">index_name</replaceable> ON <replaceable class="parameter">table</replaceable>
28 29
    [ USING <replaceable class="parameter">acc_method</replaceable> ] ( <replaceable class="parameter">func_name</replaceable>( <replaceable class="parameter">column</replaceable> [, ... ]) [ <replaceable class="parameter">ops_name</replaceable> ] )
    [ WHERE <replaceable class="parameter">predicate</replaceable> ]
30 31 32 33 34 35 36
  </synopsis>

  <refsect2 id="R2-SQL-CREATEINDEX-1">
   <refsect2info>
    <date>1998-09-09</date>
   </refsect2info>
   <title>
37
    Inputs
38 39 40 41 42 43 44 45 46 47 48
   </title>
   <para>

    <variablelist>
     <varlistentry>
      <term>UNIQUE</term>
      <listitem>
       <para>
	Causes the system to check for
	duplicate values in the table when the index is created (if data
	already exist) and each time data is added. Attempts to
Thomas G. Lockhart's avatar
Thomas G. Lockhart committed
49 50
	insert or update data which would result in duplicate entries
	will generate an error.
51 52 53
       </para>
      </listitem>
     </varlistentry>
54

55 56 57 58
     <varlistentry>
      <term><replaceable class="parameter">index_name</replaceable></term>
      <listitem>
       <para>
59 60 61
	The name of the index to be created.  No schema name can be included
	here; the index is always created in the same schema as its parent
	table.
62 63 64
       </para>
      </listitem>
     </varlistentry>
65

66 67 68 69
     <varlistentry>
      <term><replaceable class="parameter">table</replaceable></term>
      <listitem>
       <para>
70
	The name (possibly schema-qualified) of the table to be indexed.
71 72 73
       </para>
      </listitem>
     </varlistentry>
74

75
     <varlistentry>
76
      <term><replaceable class="parameter">acc_method</replaceable></term>
77 78
      <listitem>
       <para>
79 80 81 82
	The name of the access method to be used for the index. The
	default access method is <literal>BTREE</literal>.
	<application>PostgreSQL</application> provides four access
	methods for indexes:
83

84 85
	<variablelist>
	 <varlistentry>
86
	  <term><literal>BTREE</></term>
87 88
	  <listitem>
	   <para>
89
	    an implementation of Lehman-Yao
90
	    high-concurrency B-trees.
91 92 93
	   </para>
	  </listitem>
	 </varlistentry>
94

95
	 <varlistentry>
96
	  <term><literal>RTREE</></term>
97
	  <listitem>
98
	   <para>implements standard R-trees using Guttman's
99 100 101 102
	    quadratic split algorithm.
	   </para>
	  </listitem>
	 </varlistentry>
103

104
	 <varlistentry>
105
	  <term><literal>HASH</></term>
106 107 108 109 110 111
	  <listitem>
	   <para>
	    an implementation of Litwin's linear hashing.
	   </para>
	  </listitem>
	 </varlistentry>
112 113

	 <varlistentry>
114
	  <term><literal>GIST</></term>
115 116 117 118 119 120
	  <listitem>
	   <para>
	    Generalized Index Search Trees.
	   </para>
	  </listitem>
	 </varlistentry>
121 122 123 124 125 126 127 128 129 130 131 132 133
	</variablelist>
       </para>
      </listitem>
     </varlistentry>

     <varlistentry>
      <term><replaceable class="parameter">column</replaceable></term>
      <listitem>
       <para>
	The name of a column of the table.
       </para>
      </listitem>
     </varlistentry>
134

135 136 137 138
     <varlistentry>
      <term><replaceable class="parameter">ops_name</replaceable></term>
      <listitem>
       <para>
139
	An associated operator class. See below for details.
140 141 142
       </para>
      </listitem>
     </varlistentry>
143

144 145 146 147
     <varlistentry>
      <term><replaceable class="parameter">func_name</replaceable></term>
      <listitem>
       <para>
148
	A function, which returns a value that can be indexed.
149 150 151
       </para>
      </listitem>
     </varlistentry>
152 153 154 155 156 157 158 159 160

     <varlistentry>
      <term><replaceable class="parameter">predicate</replaceable></term>
      <listitem>
       <para>
	Defines the constraint expression for a partial index.
       </para>
      </listitem>
     </varlistentry>
161 162
    </variablelist>
   </para>
163 164 165 166 167 168 169
  </refsect2>

  <refsect2 id="R2-SQL-CREATEINDEX-2">
   <refsect2info>
    <date>1998-09-09</date>
   </refsect2info>
   <title>
170
    Outputs
171 172
   </title>
   <para>
173

174 175 176
    <variablelist>
     <varlistentry>
      <term><computeroutput>
177
CREATE INDEX
178 179 180 181 182 183 184
       </computeroutput></term>
      <listitem>
       <para>
	The message returned if the index is successfully created.
       </para>
      </listitem>
     </varlistentry>
185

186 187 188 189 190 191
     <varlistentry>
      <term><computeroutput>
ERROR: Cannot create index: 'index_name' already exists.
       </computeroutput></term>
      <listitem>
       <para>
192 193 194 195 196 197
	This error occurs if it is impossible to create the index.
       </para>
      </listitem>
     </varlistentry>
    </variablelist>
   </para>
198 199 200 201 202 203 204 205
  </refsect2>
 </refsynopsisdiv>

 <refsect1 id="R1-SQL-CREATEINDEX-1">
  <refsect1info>
   <date>1998-09-09</date>
  </refsect1info>
  <title>
206
   Description
207 208
  </title>
  <para>
209
   <command>CREATE INDEX</command> constructs an index 
210 211
   <replaceable class="parameter">index_name</replaceable>
   on the specified <replaceable class="parameter">table</replaceable>.
212

213 214 215 216 217 218 219
   <tip>
    <para>
     Indexes are primarily used to enhance database performance.
     But inappropriate use will result in slower performance.
    </para>
   </tip>
  </para>
220

221
  <para>
222 223 224
   In the first syntax shown above, the key field(s) for the
   index are specified as column names.
   Multiple fields can be specified if the index access method supports
225
   multicolumn indexes.
226
  </para>
227

228
  <para>
229 230 231 232 233 234 235 236 237
   In the second syntax shown above, an index is defined on the result
   of a user-specified function <replaceable
   class="parameter">func_name</replaceable> applied to one or more
   columns of a single table. These <firstterm>functional
   indexes</firstterm> can be used to obtain fast access to data based
   on operators that would normally require some transformation to apply
   them to the base data. For example, a functional index on
   <literal>upper(col)</> would allow the clause
   <literal>WHERE upper(col) = 'JIM'</> to use an index.
238 239
  </para>

240
  <para>
241 242 243 244 245 246 247 248 249
   <application>PostgreSQL</application> provides B-tree, R-tree, hash,
   and GiST access methods for indexes. The B-tree access method is an
   implementation of Lehman-Yao high-concurrency B-trees. The R-tree
   access method implements standard R-trees using Guttman's quadratic
   split algorithm. The hash access method is an implementation of
   Litwin's linear hashing. We mention the algorithms used solely to
   indicate that all of these access methods are fully dynamic and do
   not have to be optimized periodically (as is the case with, for
   example, static hash access methods).
250 251
  </para>

252 253 254 255 256 257 258 259 260
  <para>
    When the <command>WHERE</command> clause is present, a
    <firstterm>partial index</firstterm> is created.
    A partial index is an index that contains entries for only a portion of
    a table, usually a portion that is somehow more interesting than the
    rest of the table. For example, if you have a table that contains both
    billed and unbilled orders where the unbilled orders take up a small
    fraction of the total table and yet that is an often used section, you
    can improve performance by creating an index on just that portion.
261 262 263
    Another possible application is to use <command>WHERE</command> with
    <command>UNIQUE</command> to enforce uniqueness over a subset of a
    table.
264 265 266 267 268
  </para>

  <para>
    The expression used in the <command>WHERE</command> clause may refer
    only to columns of the underlying table (but it can use all columns,
269
    not only the one(s) being indexed).  Presently, subqueries and
270 271 272 273 274
    aggregate expressions are also forbidden in <command>WHERE</command>.
  </para>

  <para>
   All functions and operators used in an index definition must be
275
   <firstterm>immutable</>, that is, their results must depend only on
276 277 278
   their input arguments and never on any outside influence (such as
   the contents of another table or the current time).  This restriction
   ensures that the behavior of the index is well-defined.  To use a
279
   user-defined function in an index, remember to mark the function immutable
280
   when you create it.
281 282
  </para>

283
  <para>
284
   Use <xref linkend="sql-dropindex" endterm="sql-dropindex-title">
285 286 287
   to remove an index.
  </para>

288 289 290 291 292
  <refsect2 id="R2-SQL-CREATEINDEX-3">
   <refsect2info>
    <date>1998-09-09</date>
   </refsect2info>
   <title>
293
    Notes
294
   </title>
295 296

   <para>
297
    The <productname>PostgreSQL</productname>
298
    query optimizer will consider using a B-tree index whenever
299
    an indexed attribute is involved in a comparison using one of:
300 301 302 303 304 305 306 307 308 309 310

    <simplelist type="inline">
     <member>&lt;</member>
     <member>&lt;=</member>
     <member>=</member>
     <member>&gt;=</member>
     <member>&gt;</member>
    </simplelist>
   </para>

   <para>
311
    The <productname>PostgreSQL</productname>
312
    query optimizer will consider using an R-tree index whenever
313 314 315 316 317 318 319 320 321 322 323 324 325 326
    an indexed attribute is involved in a comparison using one of:

    <simplelist type="inline">
     <member>&lt;&lt;</member>
     <member>&amp;&lt;</member>
     <member>&amp;&gt;</member>
     <member>&gt;&gt;</member>
     <member>@</member>
     <member>~=</member>
     <member>&amp;&amp;</member>
    </simplelist>
   </para>

   <para>
327
    The <productname>PostgreSQL</productname>
328 329 330 331
    query optimizer will consider using a hash index whenever
    an indexed attribute is involved in a comparison using
    the <literal>=</literal> operator.
   </para>
332
   <para>
333
     Testing has shown PostgreSQL's hash indexes to be similar or slower
334
     than B-tree indexes, and the index size and build time for hash
335 336 337
     indexes is much worse. Hash indexes also suffer poor performance
     under high concurrency. For these reasons, hash index use is
     discouraged.
338
   </para>
339

340
   <para>
341
    Currently, only the B-tree and gist access methods support multicolumn
342
    indexes. Up to 32 keys may be specified by default (this limit
343
    can be altered when building
344
    <application>PostgreSQL</application>).  Only B-tree currently supports
345
    unique indexes.
346
   </para>
347

348 349
  <para>
   An <firstterm>operator class</firstterm> can be specified for each
350 351
   column of an index. The operator class identifies the operators to be
   used by the index for that column. For example, a B-tree index on
352 353
   four-byte integers would use the <literal>int4_ops</literal> class;
   this operator class includes comparison functions for four-byte
354 355
   integers. In practice the default operator class for the field's data
   type is usually sufficient. The main point of having operator classes
Bruce Momjian's avatar
Bruce Momjian committed
356
   is that for some data types, there could be more than one meaningful
357 358 359 360 361
   ordering. For example, we might want to sort a complex-number data
   type either by absolute value or by real part. We could do this by
   defining two operator classes for the data type and then selecting
   the proper class when making an index. There are also some operator
   classes with special purposes:
362 363 364 365 366

   <itemizedlist>
    <listitem>
     <para>
      The operator classes <literal>box_ops</literal> and
367
      <literal>bigbox_ops</literal> both support R-tree indexes on the
Bruce Momjian's avatar
Bruce Momjian committed
368
      <literal>box</literal> data type.
369
      The difference between them is that <literal>bigbox_ops</literal>
370
      scales box coordinates down, to avoid floating-point exceptions from
371
      doing multiplication, addition, and subtraction on very large
372 373 374
      floating-point coordinates.  (Note: this was true some time ago,
      but currently the two operator classes both use floating point
      and are effectively identical.)
375 376 377 378
     </para>
    </listitem>
   </itemizedlist>
  </para>
379 380

   <para>
381
    The following query shows all defined operator classes:
382 383

    <programlisting>
384
SELECT am.amname AS acc_method,
385 386 387 388
       opc.opcname AS ops_name
    FROM pg_am am, pg_opclass opc
    WHERE opc.opcamid = am.oid
    ORDER BY acc_method, ops_name;
389 390
    </programlisting>
   </para>
391
  </refsect2>
392 393
 </refsect1>

394 395
 <refsect1 id="R1-SQL-CREATEINDEX-2">
  <title>
396
   Usage
397
  </title>
398
  <para>To create a B-tree index on the field <literal>title</literal>
399
   in the table <literal>films</literal>:
400 401
  </para>
  <programlisting>
402 403
CREATE UNIQUE INDEX title_idx
    ON films (title);
404
  </programlisting>
405 406 407 408 409

<!--
<comment>
Is this example correct?
</comment>
410
  <para>
411
   To create a R-tree index on a point attribute so that we
412 413 414 415
   can efficiently use box operators on the result of the
   conversion function:
  </para>
  <programlisting>
416 417 418 419
CREATE INDEX pointloc
    ON points USING RTREE (point2box(location) box_ops);
SELECT * FROM points
    WHERE point2box(points.pointloc) = boxes.box;
420
  </programlisting>
421 422
-->

423
 </refsect1>
424
 
425 426
 <refsect1 id="R1-SQL-CREATEINDEX-3">
  <title>
427
   Compatibility
428
  </title>
429
  
430 431 432 433 434
  <refsect2 id="R2-SQL-CREATEINDEX-4">
   <refsect2info>
    <date>1998-09-09</date>
   </refsect2info>
   <title>
435
    SQL92
436 437
   </title>
   <para>
438
    CREATE INDEX is a <productname>PostgreSQL</productname> language extension.
439
   </para>
440
   <para>
441
    There is no <command>CREATE INDEX</command> command in SQL92.
442 443 444
   </para>
  </refsect2>
 </refsect1>
445
</refentry>
446 447 448 449

<!-- Keep this comment at the end of the file
Local variables:
mode: sgml
450
sgml-omittag:nil
451 452 453 454 455 456 457 458 459 460 461 462
sgml-shorttag:t
sgml-minimize-attributes:nil
sgml-always-quote-attributes:t
sgml-indent-step:1
sgml-indent-data:t
sgml-parent-document:nil
sgml-default-dtd-file:"../reference.ced"
sgml-exposed-tags:nil
sgml-local-catalogs:"/usr/lib/sgml/catalog"
sgml-local-ecat-files:nil
End:
-->