<?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>12.2. Tables and Indexes</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-intro.html" title="12.1. Introduction" /><link rel="next" href="textsearch-controls.html" title="12.3. Controlling Text Search" /></head><body><div xmlns="http://www.w3.org/TR/xhtml1/transitional" class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">12.2. Tables and Indexes</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="textsearch-intro.html" title="12.1. Introduction">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 11.5 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="textsearch-controls.html" title="12.3. Controlling Text Search">Next</a></td></tr></table><hr></hr></div><div class="sect1" id="TEXTSEARCH-TABLES"><div class="titlepage"><div><div><h2 class="title" style="clear: both">12.2. Tables and Indexes</h2></div></div></div><div class="toc"><dl class="toc"><dt><span class="sect2"><a href="textsearch-tables.html#TEXTSEARCH-TABLES-SEARCH">12.2.1. Searching a Table</a></span></dt><dt><span class="sect2"><a href="textsearch-tables.html#TEXTSEARCH-TABLES-INDEX">12.2.2. Creating Indexes</a></span></dt></dl></div><p>
The examples in the previous section illustrated full text matching using
simple constant strings. This section shows how to search table data,
optionally using indexes.
</p><div class="sect2" id="TEXTSEARCH-TABLES-SEARCH"><div class="titlepage"><div><div><h3 class="title">12.2.1. Searching a Table</h3></div></div></div><p>
It is possible to do a full text search without an index. A simple query
to print the <code class="structname">title</code> of each row that contains the word
<code class="literal">friend</code> in its <code class="structfield">body</code> field is:
</p><pre class="programlisting">
SELECT title
FROM pgweb
WHERE to_tsvector('english', body) @@ to_tsquery('english', 'friend');
</pre><p>
This will also find related words such as <code class="literal">friends</code>
and <code class="literal">friendly</code>, since all these are reduced to the same
normalized lexeme.
</p><p>
The query above specifies that the <code class="literal">english</code> configuration
is to be used to parse and normalize the strings. Alternatively we
could omit the configuration parameters:
</p><pre class="programlisting">
SELECT title
FROM pgweb
WHERE to_tsvector(body) @@ to_tsquery('friend');
</pre><p>
This query will use the configuration set by <a class="xref" href="runtime-config-client.html#GUC-DEFAULT-TEXT-SEARCH-CONFIG">default_text_search_config</a>.
</p><p>
A more complex example is to
select the ten most recent documents that contain <code class="literal">create</code> and
<code class="literal">table</code> in the <code class="structname">title</code> or <code class="structname">body</code>:
</p><pre class="programlisting">
SELECT title
FROM pgweb
WHERE to_tsvector(title || ' ' || body) @@ to_tsquery('create & table')
ORDER BY last_mod_date DESC
LIMIT 10;
</pre><p>
For clarity we omitted the <code class="function">coalesce</code> function calls
which would be needed to find rows that contain <code class="literal">NULL</code>
in one of the two fields.
</p><p>
Although these queries will work without an index, most applications
will find this approach too slow, except perhaps for occasional ad-hoc
searches. Practical use of text searching usually requires creating
an index.
</p></div><div class="sect2" id="TEXTSEARCH-TABLES-INDEX"><div class="titlepage"><div><div><h3 class="title">12.2.2. Creating Indexes</h3></div></div></div><p>
We can create a <acronym class="acronym">GIN</acronym> index (<a class="xref" href="textsearch-indexes.html" title="12.9. GIN and GiST Index Types">Section 12.9</a>) to speed up text searches:
</p><pre class="programlisting">
CREATE INDEX pgweb_idx ON pgweb USING GIN (to_tsvector('english', body));
</pre><p>
Notice that the 2-argument version of <code class="function">to_tsvector</code> is
used. Only text search functions that specify a configuration name can
be used in expression indexes (<a class="xref" href="indexes-expressional.html" title="11.7. Indexes on Expressions">Section 11.7</a>).
This is because the index contents must be unaffected by <a class="xref" href="runtime-config-client.html#GUC-DEFAULT-TEXT-SEARCH-CONFIG">default_text_search_config</a>. If they were affected, the
index contents might be inconsistent because different entries could
contain <code class="type">tsvector</code>s that were created with different text search
configurations, and there would be no way to guess which was which. It
would be impossible to dump and restore such an index correctly.
</p><p>
Because the two-argument version of <code class="function">to_tsvector</code> was
used in the index above, only a query reference that uses the 2-argument
version of <code class="function">to_tsvector</code> with the same configuration
name will use that index. That is, <code class="literal">WHERE
to_tsvector('english', body) @@ 'a & b'</code> can use the index,
but <code class="literal">WHERE to_tsvector(body) @@ 'a & b'</code> cannot.
This ensures that an index will be used only with the same configuration
used to create the index entries.
</p><p>
It is possible to set up more complex expression indexes wherein the
configuration name is specified by another column, e.g.:
</p><pre class="programlisting">
CREATE INDEX pgweb_idx ON pgweb USING GIN (to_tsvector(config_name, body));
</pre><p>
where <code class="literal">config_name</code> is a column in the <code class="literal">pgweb</code>
table. This allows mixed configurations in the same index while
recording which configuration was used for each index entry. This
would be useful, for example, if the document collection contained
documents in different languages. Again,
queries that are meant to use the index must be phrased to match, e.g.,
<code class="literal">WHERE to_tsvector(config_name, body) @@ 'a & b'</code>.
</p><p>
Indexes can even concatenate columns:
</p><pre class="programlisting">
CREATE INDEX pgweb_idx ON pgweb USING GIN (to_tsvector('english', title || ' ' || body));
</pre><p>
</p><p>
Another approach is to create a separate <code class="type">tsvector</code> column
to hold the output of <code class="function">to_tsvector</code>. This example is a
concatenation of <code class="literal">title</code> and <code class="literal">body</code>,
using <code class="function">coalesce</code> to ensure that one field will still be
indexed when the other is <code class="literal">NULL</code>:
</p><pre class="programlisting">
ALTER TABLE pgweb ADD COLUMN textsearchable_index_col tsvector;
UPDATE pgweb SET textsearchable_index_col =
to_tsvector('english', coalesce(title,'') || ' ' || coalesce(body,''));
</pre><p>
Then we create a <acronym class="acronym">GIN</acronym> index to speed up the search:
</p><pre class="programlisting">
CREATE INDEX textsearch_idx ON pgweb USING GIN (textsearchable_index_col);
</pre><p>
Now we are ready to perform a fast full text search:
</p><pre class="programlisting">
SELECT title
FROM pgweb
WHERE textsearchable_index_col @@ to_tsquery('create & table')
ORDER BY last_mod_date DESC
LIMIT 10;
</pre><p>
</p><p>
When using a separate column to store the <code class="type">tsvector</code>
representation,
it is necessary to create a trigger to keep the <code class="type">tsvector</code>
column current anytime <code class="literal">title</code> or <code class="literal">body</code> changes.
<a class="xref" href="textsearch-features.html#TEXTSEARCH-UPDATE-TRIGGERS" title="12.4.3. Triggers for Automatic Updates">Section 12.4.3</a> explains how to do that.
</p><p>
One advantage of the separate-column approach over an expression index
is that it is not necessary to explicitly specify the text search
configuration in queries in order to make use of the index. As shown
in the example above, the query can depend on
<code class="varname">default_text_search_config</code>. Another advantage is that
searches will be faster, since it will not be necessary to redo the
<code class="function">to_tsvector</code> calls to verify index matches. (This is more
important when using a GiST index than a GIN index; see <a class="xref" href="textsearch-indexes.html" title="12.9. GIN and GiST Index Types">Section 12.9</a>.) The expression-index approach is
simpler to set up, however, and it requires less disk space since the
<code class="type">tsvector</code> representation is not stored explicitly.
</p></div></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="textsearch-intro.html">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="textsearch.html">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="textsearch-controls.html">Next</a></td></tr><tr><td width="40%" align="left" valign="top">12.1. Introduction </td><td width="20%" align="center"><a accesskey="h" href="index.html">Home</a></td><td width="40%" align="right" valign="top"> 12.3. Controlling Text Search</td></tr></table></div></body></html>
