1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
|
<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<!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>12.1. Introduction</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="textsearch.html" title="Chapter 12. Full Text Search" /><link rel="next" href="textsearch-tables.html" title="12.2. Tables and 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">12.1. Introduction</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="textsearch.html" title="Chapter 12. Full Text Search">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="textsearch.html" title="Chapter 12. Full Text Search">Up</a></td><th width="60%" align="center">Chapter 12. Full Text Search</th><td width="10%" align="right"><a accesskey="h" href="index.html" title="PostgreSQL 15.7 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="textsearch-tables.html" title="12.2. Tables and Indexes">Next</a></td></tr></table><hr /></div><div class="sect1" id="TEXTSEARCH-INTRO"><div class="titlepage"><div><div><h2 class="title" style="clear: both">12.1. Introduction</h2></div></div></div><div class="toc"><dl class="toc"><dt><span class="sect2"><a href="textsearch-intro.html#TEXTSEARCH-DOCUMENT">12.1.1. What Is a Document?</a></span></dt><dt><span class="sect2"><a href="textsearch-intro.html#TEXTSEARCH-MATCHING">12.1.2. Basic Text Matching</a></span></dt><dt><span class="sect2"><a href="textsearch-intro.html#TEXTSEARCH-INTRO-CONFIGURATIONS">12.1.3. Configurations</a></span></dt></dl></div><p>
Full Text Searching (or just <em class="firstterm">text search</em>) provides
the capability to identify natural-language <em class="firstterm">documents</em> that
satisfy a <em class="firstterm">query</em>, and optionally to sort them by
relevance to the query. The most common type of search
is to find all documents containing given <em class="firstterm">query terms</em>
and return them in order of their <em class="firstterm">similarity</em> to the
query. Notions of <code class="varname">query</code> and
<code class="varname">similarity</code> are very flexible and depend on the specific
application. The simplest search considers <code class="varname">query</code> as a
set of words and <code class="varname">similarity</code> as the frequency of query
words in the document.
</p><p>
Textual search operators have existed in databases for years.
<span class="productname">PostgreSQL</span> has
<code class="literal">~</code>, <code class="literal">~*</code>, <code class="literal">LIKE</code>, and
<code class="literal">ILIKE</code> operators for textual data types, but they lack
many essential properties required by modern information systems:
</p><div class="itemizedlist"><ul class="itemizedlist compact" style="list-style-type: bullet; "><li class="listitem" style="list-style-type: disc"><p>
There is no linguistic support, even for English. Regular expressions
are not sufficient because they cannot easily handle derived words, e.g.,
<code class="literal">satisfies</code> and <code class="literal">satisfy</code>. You might
miss documents that contain <code class="literal">satisfies</code>, although you
probably would like to find them when searching for
<code class="literal">satisfy</code>. It is possible to use <code class="literal">OR</code>
to search for multiple derived forms, but this is tedious and error-prone
(some words can have several thousand derivatives).
</p></li><li class="listitem" style="list-style-type: disc"><p>
They provide no ordering (ranking) of search results, which makes them
ineffective when thousands of matching documents are found.
</p></li><li class="listitem" style="list-style-type: disc"><p>
They tend to be slow because there is no index support, so they must
process all documents for every search.
</p></li></ul></div><p>
Full text indexing allows documents to be <span class="emphasis"><em>preprocessed</em></span>
and an index saved for later rapid searching. Preprocessing includes:
</p><div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: none; "><li class="listitem" style="list-style-type: none"><p>
<span class="emphasis"><em>Parsing documents into <em class="firstterm">tokens</em></em></span>. It is
useful to identify various classes of tokens, e.g., numbers, words,
complex words, email addresses, so that they can be processed
differently. In principle token classes depend on the specific
application, but for most purposes it is adequate to use a predefined
set of classes.
<span class="productname">PostgreSQL</span> uses a <em class="firstterm">parser</em> to
perform this step. A standard parser is provided, and custom parsers
can be created for specific needs.
</p></li><li class="listitem" style="list-style-type: none"><p>
<span class="emphasis"><em>Converting tokens into <em class="firstterm">lexemes</em></em></span>.
A lexeme is a string, just like a token, but it has been
<em class="firstterm">normalized</em> so that different forms of the same word
are made alike. For example, normalization almost always includes
folding upper-case letters to lower-case, and often involves removal
of suffixes (such as <code class="literal">s</code> or <code class="literal">es</code> in English).
This allows searches to find variant forms of the
same word, without tediously entering all the possible variants.
Also, this step typically eliminates <em class="firstterm">stop words</em>, which
are words that are so common that they are useless for searching.
(In short, then, tokens are raw fragments of the document text, while
lexemes are words that are believed useful for indexing and searching.)
<span class="productname">PostgreSQL</span> uses <em class="firstterm">dictionaries</em> to
perform this step. Various standard dictionaries are provided, and
custom ones can be created for specific needs.
</p></li><li class="listitem" style="list-style-type: none"><p>
<span class="emphasis"><em>Storing preprocessed documents optimized for
searching</em></span>. For example, each document can be represented
as a sorted array of normalized lexemes. Along with the lexemes it is
often desirable to store positional information to use for
<em class="firstterm">proximity ranking</em>, so that a document that
contains a more <span class="quote">“<span class="quote">dense</span>”</span> region of query words is
assigned a higher rank than one with scattered query words.
</p></li></ul></div><p>
Dictionaries allow fine-grained control over how tokens are normalized.
With appropriate dictionaries, you can:
</p><div class="itemizedlist"><ul class="itemizedlist compact" style="list-style-type: bullet; "><li class="listitem" style="list-style-type: disc"><p>
Define stop words that should not be indexed.
</p></li><li class="listitem" style="list-style-type: disc"><p>
Map synonyms to a single word using <span class="application">Ispell</span>.
</p></li><li class="listitem" style="list-style-type: disc"><p>
Map phrases to a single word using a thesaurus.
</p></li><li class="listitem" style="list-style-type: disc"><p>
Map different variations of a word to a canonical form using
an <span class="application">Ispell</span> dictionary.
</p></li><li class="listitem" style="list-style-type: disc"><p>
Map different variations of a word to a canonical form using
<span class="application">Snowball</span> stemmer rules.
</p></li></ul></div><p>
A data type <code class="type">tsvector</code> is provided for storing preprocessed
documents, along with a type <code class="type">tsquery</code> for representing processed
queries (<a class="xref" href="datatype-textsearch.html" title="8.11. Text Search Types">Section 8.11</a>). There are many
functions and operators available for these data types
(<a class="xref" href="functions-textsearch.html" title="9.13. Text Search Functions and Operators">Section 9.13</a>), the most important of which is
the match operator <code class="literal">@@</code>, which we introduce in
<a class="xref" href="textsearch-intro.html#TEXTSEARCH-MATCHING" title="12.1.2. Basic Text Matching">Section 12.1.2</a>. Full text searches can be accelerated
using indexes (<a class="xref" href="textsearch-indexes.html" title="12.9. Preferred Index Types for Text Search">Section 12.9</a>).
</p><div class="sect2" id="TEXTSEARCH-DOCUMENT"><div class="titlepage"><div><div><h3 class="title">12.1.1. What Is a Document?</h3></div></div></div><a id="id-1.5.11.4.10.2" class="indexterm"></a><p>
A <em class="firstterm">document</em> is the unit of searching in a full text search
system; for example, a magazine article or email message. The text search
engine must be able to parse documents and store associations of lexemes
(key words) with their parent document. Later, these associations are
used to search for documents that contain query words.
</p><p>
For searches within <span class="productname">PostgreSQL</span>,
a document is normally a textual field within a row of a database table,
or possibly a combination (concatenation) of such fields, perhaps stored
in several tables or obtained dynamically. In other words, a document can
be constructed from different parts for indexing and it might not be
stored anywhere as a whole. For example:
</p><pre class="programlisting">
SELECT title || ' ' || author || ' ' || abstract || ' ' || body AS document
FROM messages
WHERE mid = 12;
SELECT m.title || ' ' || m.author || ' ' || m.abstract || ' ' || d.body AS document
FROM messages m, docs d
WHERE m.mid = d.did AND m.mid = 12;
</pre><p>
</p><div class="note"><h3 class="title">Note</h3><p>
Actually, in these example queries, <code class="function">coalesce</code>
should be used to prevent a single <code class="literal">NULL</code> attribute from
causing a <code class="literal">NULL</code> result for the whole document.
</p></div><p>
Another possibility is to store the documents as simple text files in the
file system. In this case, the database can be used to store the full text
index and to execute searches, and some unique identifier can be used to
retrieve the document from the file system. However, retrieving files
from outside the database requires superuser permissions or special
function support, so this is usually less convenient than keeping all
the data inside <span class="productname">PostgreSQL</span>. Also, keeping
everything inside the database allows easy access
to document metadata to assist in indexing and display.
</p><p>
For text search purposes, each document must be reduced to the
preprocessed <code class="type">tsvector</code> format. Searching and ranking
are performed entirely on the <code class="type">tsvector</code> representation
of a document — the original text need only be retrieved
when the document has been selected for display to a user.
We therefore often speak of the <code class="type">tsvector</code> as being the
document, but of course it is only a compact representation of
the full document.
</p></div><div class="sect2" id="TEXTSEARCH-MATCHING"><div class="titlepage"><div><div><h3 class="title">12.1.2. Basic Text Matching</h3></div></div></div><p>
Full text searching in <span class="productname">PostgreSQL</span> is based on
the match operator <code class="literal">@@</code>, which returns
<code class="literal">true</code> if a <code class="type">tsvector</code>
(document) matches a <code class="type">tsquery</code> (query).
It doesn't matter which data type is written first:
</p><pre class="programlisting">
SELECT 'a fat cat sat on a mat and ate a fat rat'::tsvector @@ 'cat & rat'::tsquery;
?column?
----------
t
SELECT 'fat & cow'::tsquery @@ 'a fat cat sat on a mat and ate a fat rat'::tsvector;
?column?
----------
f
</pre><p>
</p><p>
As the above example suggests, a <code class="type">tsquery</code> is not just raw
text, any more than a <code class="type">tsvector</code> is. A <code class="type">tsquery</code>
contains search terms, which must be already-normalized lexemes, and
may combine multiple terms using AND, OR, NOT, and FOLLOWED BY operators.
(For syntax details see <a class="xref" href="datatype-textsearch.html#DATATYPE-TSQUERY" title="8.11.2. tsquery">Section 8.11.2</a>.) There are
functions <code class="function">to_tsquery</code>, <code class="function">plainto_tsquery</code>,
and <code class="function">phraseto_tsquery</code>
that are helpful in converting user-written text into a proper
<code class="type">tsquery</code>, primarily by normalizing words appearing in
the text. Similarly, <code class="function">to_tsvector</code> is used to parse and
normalize a document string. So in practice a text search match would
look more like this:
</p><pre class="programlisting">
SELECT to_tsvector('fat cats ate fat rats') @@ to_tsquery('fat & rat');
?column?
----------
t
</pre><p>
Observe that this match would not succeed if written as
</p><pre class="programlisting">
SELECT 'fat cats ate fat rats'::tsvector @@ to_tsquery('fat & rat');
?column?
----------
f
</pre><p>
since here no normalization of the word <code class="literal">rats</code> will occur.
The elements of a <code class="type">tsvector</code> are lexemes, which are assumed
already normalized, so <code class="literal">rats</code> does not match <code class="literal">rat</code>.
</p><p>
The <code class="literal">@@</code> operator also
supports <code class="type">text</code> input, allowing explicit conversion of a text
string to <code class="type">tsvector</code> or <code class="type">tsquery</code> to be skipped
in simple cases. The variants available are:
</p><pre class="programlisting">
tsvector @@ tsquery
tsquery @@ tsvector
text @@ tsquery
text @@ text
</pre><p>
</p><p>
The first two of these we saw already.
The form <code class="type">text</code> <code class="literal">@@</code> <code class="type">tsquery</code>
is equivalent to <code class="literal">to_tsvector(x) @@ y</code>.
The form <code class="type">text</code> <code class="literal">@@</code> <code class="type">text</code>
is equivalent to <code class="literal">to_tsvector(x) @@ plainto_tsquery(y)</code>.
</p><p>
Within a <code class="type">tsquery</code>, the <code class="literal">&</code> (AND) operator
specifies that both its arguments must appear in the document to have a
match. Similarly, the <code class="literal">|</code> (OR) operator specifies that
at least one of its arguments must appear, while the <code class="literal">!</code> (NOT)
operator specifies that its argument must <span class="emphasis"><em>not</em></span> appear in
order to have a match.
For example, the query <code class="literal">fat & ! rat</code> matches documents that
contain <code class="literal">fat</code> but not <code class="literal">rat</code>.
</p><p>
Searching for phrases is possible with the help of
the <code class="literal"><-></code> (FOLLOWED BY) <code class="type">tsquery</code> operator, which
matches only if its arguments have matches that are adjacent and in the
given order. For example:
</p><pre class="programlisting">
SELECT to_tsvector('fatal error') @@ to_tsquery('fatal <-> error');
?column?
----------
t
SELECT to_tsvector('error is not fatal') @@ to_tsquery('fatal <-> error');
?column?
----------
f
</pre><p>
There is a more general version of the FOLLOWED BY operator having the
form <code class="literal"><<em class="replaceable"><code>N</code></em>></code>,
where <em class="replaceable"><code>N</code></em> is an integer standing for the difference between
the positions of the matching lexemes. <code class="literal"><1></code> is
the same as <code class="literal"><-></code>, while <code class="literal"><2></code>
allows exactly one other lexeme to appear between the matches, and so
on. The <code class="literal">phraseto_tsquery</code> function makes use of this
operator to construct a <code class="literal">tsquery</code> that can match a multi-word
phrase when some of the words are stop words. For example:
</p><pre class="programlisting">
SELECT phraseto_tsquery('cats ate rats');
phraseto_tsquery
-------------------------------
'cat' <-> 'ate' <-> 'rat'
SELECT phraseto_tsquery('the cats ate the rats');
phraseto_tsquery
-------------------------------
'cat' <-> 'ate' <2> 'rat'
</pre><p>
</p><p>
A special case that's sometimes useful is that <code class="literal"><0></code>
can be used to require that two patterns match the same word.
</p><p>
Parentheses can be used to control nesting of the <code class="type">tsquery</code>
operators. Without parentheses, <code class="literal">|</code> binds least tightly,
then <code class="literal">&</code>, then <code class="literal"><-></code>,
and <code class="literal">!</code> most tightly.
</p><p>
It's worth noticing that the AND/OR/NOT operators mean something subtly
different when they are within the arguments of a FOLLOWED BY operator
than when they are not, because within FOLLOWED BY the exact position of
the match is significant. For example, normally <code class="literal">!x</code> matches
only documents that do not contain <code class="literal">x</code> anywhere.
But <code class="literal">!x <-> y</code> matches <code class="literal">y</code> if it is not
immediately after an <code class="literal">x</code>; an occurrence of <code class="literal">x</code>
elsewhere in the document does not prevent a match. Another example is
that <code class="literal">x & y</code> normally only requires that <code class="literal">x</code>
and <code class="literal">y</code> both appear somewhere in the document, but
<code class="literal">(x & y) <-> z</code> requires <code class="literal">x</code>
and <code class="literal">y</code> to match at the same place, immediately before
a <code class="literal">z</code>. Thus this query behaves differently from
<code class="literal">x <-> z & y <-> z</code>, which will match a
document containing two separate sequences <code class="literal">x z</code> and
<code class="literal">y z</code>. (This specific query is useless as written,
since <code class="literal">x</code> and <code class="literal">y</code> could not match at the same place;
but with more complex situations such as prefix-match patterns, a query
of this form could be useful.)
</p></div><div class="sect2" id="TEXTSEARCH-INTRO-CONFIGURATIONS"><div class="titlepage"><div><div><h3 class="title">12.1.3. Configurations</h3></div></div></div><p>
The above are all simple text search examples. As mentioned before, full
text search functionality includes the ability to do many more things:
skip indexing certain words (stop words), process synonyms, and use
sophisticated parsing, e.g., parse based on more than just white space.
This functionality is controlled by <em class="firstterm">text search
configurations</em>. <span class="productname">PostgreSQL</span> comes with predefined
configurations for many languages, and you can easily create your own
configurations. (<span class="application">psql</span>'s <code class="command">\dF</code> command
shows all available configurations.)
</p><p>
During installation an appropriate configuration is selected and
<a class="xref" href="runtime-config-client.html#GUC-DEFAULT-TEXT-SEARCH-CONFIG">default_text_search_config</a> is set accordingly
in <code class="filename">postgresql.conf</code>. If you are using the same text search
configuration for the entire cluster you can use the value in
<code class="filename">postgresql.conf</code>. To use different configurations
throughout the cluster but the same configuration within any one database,
use <code class="command">ALTER DATABASE ... SET</code>. Otherwise, you can set
<code class="varname">default_text_search_config</code> in each session.
</p><p>
Each text search function that depends on a configuration has an optional
<code class="type">regconfig</code> argument, so that the configuration to use can be
specified explicitly. <code class="varname">default_text_search_config</code>
is used only when this argument is omitted.
</p><p>
To make it easier to build custom text search configurations, a
configuration is built up from simpler database objects.
<span class="productname">PostgreSQL</span>'s text search facility provides
four types of configuration-related database objects:
</p><div class="itemizedlist"><ul class="itemizedlist compact" style="list-style-type: bullet; "><li class="listitem" style="list-style-type: disc"><p>
<em class="firstterm">Text search parsers</em> break documents into tokens
and classify each token (for example, as words or numbers).
</p></li><li class="listitem" style="list-style-type: disc"><p>
<em class="firstterm">Text search dictionaries</em> convert tokens to normalized
form and reject stop words.
</p></li><li class="listitem" style="list-style-type: disc"><p>
<em class="firstterm">Text search templates</em> provide the functions underlying
dictionaries. (A dictionary simply specifies a template and a set
of parameters for the template.)
</p></li><li class="listitem" style="list-style-type: disc"><p>
<em class="firstterm">Text search configurations</em> select a parser and a set
of dictionaries to use to normalize the tokens produced by the parser.
</p></li></ul></div><p>
Text search parsers and templates are built from low-level C functions;
therefore it requires C programming ability to develop new ones, and
superuser privileges to install one into a database. (There are examples
of add-on parsers and templates in the <code class="filename">contrib/</code> area of the
<span class="productname">PostgreSQL</span> distribution.) Since dictionaries and
configurations just parameterize and connect together some underlying
parsers and templates, no special privilege is needed to create a new
dictionary or configuration. Examples of creating custom dictionaries and
configurations appear later in this chapter.
</p></div></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="textsearch.html" title="Chapter 12. Full Text Search">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="textsearch.html" title="Chapter 12. Full Text Search">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="textsearch-tables.html" title="12.2. Tables and Indexes">Next</a></td></tr><tr><td width="40%" align="left" valign="top">Chapter 12. Full Text Search </td><td width="20%" align="center"><a accesskey="h" href="index.html" title="PostgreSQL 15.7 Documentation">Home</a></td><td width="40%" align="right" valign="top"> 12.2. Tables and Indexes</td></tr></table></div></body></html>
|
