<?xml version="1.0" encoding="ISO-8859-1"?>
<!-- 
 -  
 -  This file is part of the OpenLink Software Virtuoso Open-Source (VOS)
 -  project.
 -  
 -  Copyright (C) 1998-2018 OpenLink Software
 -  
 -  This project is free software; you can redistribute it and/or modify it
 -  under the terms of the GNU General Public License as published by the
 -  Free Software Foundation; only version 2 of the License, dated June 1991.
 -  
 -  This program is distributed in the hope that it will be useful, but
 -  WITHOUT ANY WARRANTY; without even the implied warranty of
 -  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
 -  General Public License for more details.
 -  
 -  You should have received a copy of the GNU General Public License along
 -  with this program; if not, write to the Free Software Foundation, Inc.,
 -  51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
 -  
 -  
-->
<refentry id="fn_vt_create_text_index">
  <refmeta>
    <refentrytitle>vt_create_text_index</refentrytitle>
    <refmiscinfo>ft</refmiscinfo>
  </refmeta>
  <refnamediv>
    <refname>vt_create_text_index</refname>
    <refpurpose>Add text index to an existing table</refpurpose>
  </refnamediv>
  <refsynopsisdiv>
    <funcsynopsis id="fsyn_vt_create_text_index"> 
      <funcprototype id="fproto_vt_create_text_index">
        <funcdef><function>vt_create_text_index</function></funcdef>
        <paramdef>in <parameter>table</parameter> varchar</paramdef>
        <paramdef>in <parameter>col</parameter> varchar</paramdef>
        <paramdef>in <parameter>id_col</parameter> varchar</paramdef>
        <paramdef>in <parameter>is_xml</parameter> integer</paramdef>
        <paramdef>in <parameter>defer_generation</parameter> integer</paramdef>
        <paramdef>in <parameter>clustered_columns</parameter> any</paramdef>
        <paramdef>in <parameter>use_hook_function</parameter> integer</paramdef>
        <paramdef><optional>in <parameter>language_name</parameter> varchar</optional></paramdef>
        <paramdef><optional>in <parameter>encoding_name</parameter> integer</optional></paramdef>
        <paramdef><optional>in <parameter>silent</parameter> integer</optional></paramdef>        
      </funcprototype>
    </funcsynopsis>
  </refsynopsisdiv>
  <refsect1 id="desc"><title>Description</title>
    <para>
The vt_create_text_index procedure adds a text index to an existing table.
There can at most be one text index per table, including super tables and
subtables.
  </para>
    <para>
The table argument is a string naming the table. The column is the name of the column
to index. The id_col should be the name of a unique integer row identifier column.  If null,
the system will either add such a column or use an existing integer primary key if one
is available.  The is_xml argument, if non-0, specifies that the values of the indexed column
should be checked for XML well formedness and that the XML structure should be taken into account in indexing the
values.
  </para>
    <para>
Use the CREATE TEXT INDEX statement as an alternative to the vt_create_text_index function.
  </para>
				<para>
In order for a table to be referenced in a text index it must have
a uniquely identifying integer key.  If the table in question has such a 
key this can be used as the id column.  If there is no such column this 
procedure makes one.  Using a previously existing identifier column saves 
space and if that is the primary key of the table this also saves in 
retrieval time.
	</para>
			<para>
If the table being indexed has a single part integer primary key 
vt_create_text_index will automatically use this as the identifier.  Note 
that the zero and negative numbers may not be used as identifier values.
	</para>
			<para>
Creating the index will read through the table&apos;s contents and generate 
the index.  When the table is changed the index can either be updated after 
each change or periodically, depending on the application needs.  The rationale 
for background maintenance of the text index is that it is up to several times
more efficient to maintain a text index in batches of several changed documents 
than after each single document change.  The default maintenance mode is 
synchronous, meaning that each insert, delete or update of the indexed column 
will be immediately reflected in the index.  This mode can be set using 
the <function>vt_batch_update()</function> procedure.
The mode should be set to batch if there are 
any massive operations on the table.
	</para>
			<para>
It will create two additional tables:
	</para>
			<screen>
 &lt;datatable&gt;_&lt;datacolumn&gt;_WORDS
	</screen>
			<para>
 and
	</para>
			<screen>
VTLOG_&lt;datatable_qualifieri&gt;_&lt;data_table_owner&gt;_&lt;data_table_name&gt;
	</screen>
			<para>
 and two procedures:
	</para>
			<screen>
 vt_index_&lt;datatable_qualifier&gt;_&lt;data_table_owner&gt;_&lt;data_table_name&gt; 
		(in to_delete integer)
	</screen>
			<para>
 and
	</para>
			<screen>
 vt_inc_index&lt;datatable_qualifieri&gt;_&lt;data_table_owner&gt;_&lt;data_table_name&gt; ().
</screen>
			<para>
The table &lt;datatable&gt;_&lt;datafield&gt;_WORDS contains the full text index data.
</para>
			<para>
The table VTLOG_&lt;datatable&gt;_&lt;datafield&gt; is an update tracking table, similar to the snapshot log table but using the key column instead of the primary key.
</para>

    <note><title>Note:</title>
      <para>in the transaction semantics section that there is the sync mode for the purpose of creating a text index.</para>
	</note>
  </refsect1>
  <refsect1 id="params"><title>Parameters</title>
    <refsect2><title>table</title>
      <para>the table containing the data to index.</para></refsect2>
    <refsect2><title>col</title>
      <para>the column in the data table containing the data to index (a long varchar column).</para></refsect2>
    <refsect2><title>id_col</title>
      <para>an integer unique indexed column used by the free text index as a key.  If Virtuoso tries  
to choose such a column among the existing in the table and if it doesn&apos;t find a suitable column it 
adds such a column with the name &lt;datacolumn&gt;_ID, fills that up and makes an unique index on it.</para></refsect2>
    <refsect2><title>is_xml</title>
      <para>if greater than 0 installs two additional triggers before insert and before update on the data table 
to ensure the data being inserted into it are valid XML documents.</para></refsect2>
    <refsect2><title>defer_generation</title>
      <para>if nonzero then the free-text index will not be filled by actual data
immediately after the creation. It will remain empty until explicit request for
&quot;incremental indexing&quot;.</para></refsect2>
    <refsect2><title>clustered_columns</title>
      <para>a vector of names of &quot;clustered columns&quot; or NULL to not store such data in the index.</para></refsect2>
    <refsect2><title>use_hook_function</title>
      <para>if nonzero, two user-defined Virtuoso/PL functions will be
called when free-text data are updated, not the default system routine.
These functions are recognized by their special names:
&lt;datatable&gt;_&lt;datafield&gt;_INDEX_HOOK will be called to index
new documents and
&lt;datatable&gt;_&lt;datafield&gt;_UNINDEX_HOOK will be called to
remove obsolete index information about deleted documents.</para></refsect2>
    <refsect2><title>language_name</title>
      <para>the name of the language that is used for building the index.
If the parameter is omitted or is equal to '*ini*' string, indexing routines
will use the language specified in server's configuration.</para></refsect2>
    <refsect2><title>encoding_name</title>
      <para>the name of the encoding that is used by default to
index source texts.
If the parameter is omitted or is equal to '*ini*' string, indexing routines
will use the encoding specified by charset of the RDBMS connection that is
in use when the index is created.</para></refsect2>
    <refsect2><title>silent</title>
      <para>Flag with values 1/0. Default is 0. When set to 0, the function signals error if the text index exists. 
      	If set to 1, the function signals no error if the text index exists.</para></refsect2>
  </refsect1>
<!--
  <refsect1 id="errors"><title>Errors</title>
    <table><title>Errors signalled by</title>
      <tgroup cols="4">
       <thead><row><entry>SQLState</entry><entry>Error Code</entry><entry>Error Text</entry><entry>Description</entry></row></thead>
	<tbody>
	  <row>
	    <entry><errorcode></errorcode></entry>
	    <entry><errorcode></errorcode></entry>
	    <entry><errorname></errorname></entry>
	    <entry></entry>
	  </row>
	</tbody>
      </tgroup>
    </table>
  </refsect1>
  -->
  <refsect1 id="examples"><title>Examples</title>
			<example id="ex_vt_create_text_index">
				<title>Creating A Text Index Using FT functions</title>
				<programlisting>
create table xml_text (xt_id integer, xt_file varchar,
			xt_text long varchar, primary key (xt_id));

create index xt_file on xml_text (xt_file);

vt_create_text_index (&apos;xml_text&apos;,
			&apos;xt_text&apos;, null, 1);
</programlisting>
				<para>-- create a text index on xt_text with XML well formedness checking on</para>
				<programlisting>
VT_INDEX_DB_DBA_XML_TEXT (0);
</programlisting>
				<para>-- populate the text index based in the data in the table.</para>
				<programlisting>
VT_inc_INDEX_DB_DBA_XML_TEXT ();
</programlisting>
				<para>
-- refresh the index to reflect changes to the data since the last call
</para>
				<para>-- to this function or the initial call to VT_INDEX_DB_DBA_XML_TEXT.
</para>
				<programlisting>
vt_batch_update (&apos;DB.DBA.XML_TEXT&apos;, &apos;on&apos;, 10);
</programlisting>
				<para>-- sets the index maintenance to be asynchronous, refreshed at maximum intervals
of 10 minutes or at every scheduler cycle if the latter is longer than 10 minutes.
</para>
				<programlisting>
vt_batch_update (&apos;DB.DBA.XML_TEXT&apos;, &apos;off&apos;, null);
</programlisting>
				<para>
-- sets the index maintenance to be synchronous, within the same transaction
</para>
				<para>
-- as the operation changing the indexed data.
</para>
			</example>
  </refsect1>
<!--  <refsect1 id="seealso"><title>See Also</title>
    <para><link linkend="fn"></link></para>
  </refsect1> -->
</refentry>