Minor improvements to the trigger documentation, and a few SGML fixes.

This commit is contained in:
Neil Conway 2004-01-22 19:50:21 +00:00
parent 5ad7d65da4
commit 58ae3cf12c
2 changed files with 74 additions and 44 deletions

View File

@ -1,5 +1,5 @@
<!-- <!--
$PostgreSQL: pgsql/doc/src/sgml/plpgsql.sgml,v 1.32 2003/11/30 05:45:22 momjian Exp $ $PostgreSQL: pgsql/doc/src/sgml/plpgsql.sgml,v 1.33 2004/01/22 19:50:21 neilc Exp $
--> -->
<chapter id="plpgsql"> <chapter id="plpgsql">
@ -729,7 +729,7 @@ RENAME <replaceable>oldname</replaceable> TO <replaceable>newname</replaceable>;
<para> <para>
Using the <literal>RENAME</literal> declaration you can change the Using the <literal>RENAME</literal> declaration you can change the
name of a variable, record or row. This is primarily useful if name of a variable, record or row. This is primarily useful if
<literal>NEW</literal> or <literal>OLD</literal> should be <varname>NEW</varname> or <varname>OLD</varname> should be
referenced by another name inside a trigger procedure. See also referenced by another name inside a trigger procedure. See also
<literal>ALIAS</literal>. <literal>ALIAS</literal>.
</para> </para>
@ -2176,7 +2176,7 @@ RAISE EXCEPTION ''Inexistent ID --> %'', user_id;
<para> <para>
Data type <type>RECORD</type>; variable holding the new Data type <type>RECORD</type>; variable holding the new
database row for <command>INSERT</>/<command>UPDATE</> operations in row-level database row for <command>INSERT</>/<command>UPDATE</> operations in row-level
triggers. This variable is null in statement-level triggers. triggers. This variable is <symbol>NULL</symbol> in statement-level triggers.
</para> </para>
</listitem> </listitem>
</varlistentry> </varlistentry>
@ -2187,7 +2187,7 @@ RAISE EXCEPTION ''Inexistent ID --> %'', user_id;
<para> <para>
Data type <type>RECORD</type>; variable holding the old Data type <type>RECORD</type>; variable holding the old
database row for <command>UPDATE</>/<command>DELETE</> operations in row-level database row for <command>UPDATE</>/<command>DELETE</> operations in row-level
triggers. This variable is null in statement-level triggers. triggers. This variable is <symbol>NULL</symbol> in statement-level triggers.
</para> </para>
</listitem> </listitem>
</varlistentry> </varlistentry>
@ -2281,9 +2281,9 @@ RAISE EXCEPTION ''Inexistent ID --> %'', user_id;
</para> </para>
<para> <para>
A trigger function must return either null or a record/row value A trigger function must return either <symbol>NULL</symbol> or a
having exactly the structure of the table the trigger was fired record/row value having exactly the structure of the table the
for. trigger was fired for.
</para> </para>
<para> <para>

View File

@ -1,5 +1,5 @@
<!-- <!--
$PostgreSQL: pgsql/doc/src/sgml/trigger.sgml,v 1.33 2003/11/29 19:51:38 pgsql Exp $ $PostgreSQL: pgsql/doc/src/sgml/trigger.sgml,v 1.34 2004/01/22 19:50:21 neilc Exp $
--> -->
<chapter id="triggers"> <chapter id="triggers">
@ -45,50 +45,69 @@ $PostgreSQL: pgsql/doc/src/sgml/trigger.sgml,v 1.33 2003/11/29 19:51:38 pgsql Ex
</para> </para>
<para> <para>
Trigger functions return a table row (a value of type There are two types of triggers: per-row triggers and
<structname>HeapTuple</>) to the calling executor. per-statement triggers. In a per-row trigger, the trigger function
A trigger fired before an operation has the following choices: is invoked once for every row that is affected by the statement
that fired the trigger. In contrast, a per-statement trigger is
invoked only once when an appropriate statement is executed,
regardless of the number of rows affected by that statement. In
particular, a statement that affects zero rows will still result
in the execution of any applicable per-statement triggers. These
two types of triggers are sometimes called <quote>row-level
triggers</quote> and <quote>statement-level triggers</quote>,
respectively.
</para>
<para>
Trigger functions invoked by per-statement triggers should always
return <symbol>NULL</symbol>. Trigger functions invoked by per-row
triggers can return a table row (a value of
type <structname>HeapTuple</structname>) to the calling executor,
if they choose. A row-level trigger fired before an operation has
the following choices:
<itemizedlist> <itemizedlist>
<listitem> <listitem>
<para> <para>
It can return a <symbol>NULL</> pointer to skip the operation It can return <symbol>NULL</> to skip the operation for the
for the current row (and so the row will not be current row. This instructs the executor to not perform the
inserted/updated/deleted). row-level operation that invoked the trigger (the insertion or
modification of a particular table row).
</para> </para>
</listitem> </listitem>
<listitem> <listitem>
<para> <para>
For <command>INSERT</command> and <command>UPDATE</command> For row-level <command>INSERT</command>
triggers only, the returned row becomes the row that will and <command>UPDATE</command> triggers only, the returned row
be inserted or will replace the row being updated. This becomes the row that will be inserted or will replace the row
allows the trigger function to modify the row being inserted or being updated. This allows the trigger function to modify the
updated. row being inserted or updated.
</para> </para>
</listitem> </listitem>
</itemizedlist> </itemizedlist>
A before trigger that does not intend to cause either of these A row-level before trigger that does not intend to cause either of
behaviors must be careful to return as its result the same row that was these behaviors must be careful to return as its result the same
passed in (that is, the NEW row for <command>INSERT</command> and row that was passed in (that is, the <varname>NEW</varname> row
<command>UPDATE</command> triggers, the OLD row for for <command>INSERT</command> and <command>UPDATE</command>
triggers, the <varname>OLD</varname> row for
<command>DELETE</command> triggers). <command>DELETE</command> triggers).
</para> </para>
<para> <para>
The return The return value is ignored for row-level triggers fired after an
value is ignored for triggers fired after an operation, and so operation, and so they may as well return <symbol>NULL</>.
they may as well return <symbol>NULL</>.
</para> </para>
<para> <para>
If more than one trigger is defined for the same event on the same If more than one trigger is defined for the same event on the same
relation, the triggers will be fired in alphabetical order by trigger relation, the triggers will be fired in alphabetical order by
name. In the case of before triggers, the possibly-modified row trigger name. In the case of before triggers, the
returned by each trigger becomes the input to the next trigger. possibly-modified row returned by each trigger becomes the input
If any before trigger returns a <symbol>NULL</> pointer, the to the next trigger. If any before trigger returns
operation is abandoned and subsequent triggers are not fired. <symbol>NULL</>, the operation is abandoned and subsequent
triggers are not fired.
</para> </para>
<para> <para>
@ -134,30 +153,41 @@ $PostgreSQL: pgsql/doc/src/sgml/trigger.sgml,v 1.33 2003/11/29 19:51:38 pgsql Ex
is fired for. Briefly: is fired for. Briefly:
<itemizedlist> <itemizedlist>
<listitem> <listitem>
<para> <para>
The data change (insertion, update, or deletion) causing the trigger Statement-level triggers follow simple visibility rules: none of
to fire is naturally the changes made by a statement are visible to statement-level
<emphasis>not</emphasis> visible to SQL commands executed in a triggers that are invoked before the statement, whereas all
before trigger, because it hasn't happened yet. modifications are visible to statement-level after triggers.
</para> </para>
</listitem> </listitem>
<listitem> <listitem>
<para> <para>
However, SQL commands executed in a before trigger The data change (insertion, update, or deletion) causing the
<emphasis>will</emphasis> see the effects of data changes trigger to fire is naturally <emphasis>not</emphasis> visible
for rows previously processed in the same outer command. This to SQL commands executed in a row-level before trigger, because
requires caution, since the ordering of these change events it hasn't happened yet.
is not in general predictable; a SQL command that affects
multiple rows may visit the rows in any order.
</para> </para>
</listitem> </listitem>
<listitem> <listitem>
<para> <para>
When an after trigger is fired, all data changes made by the outer However, SQL commands executed in a row-level before
command are already complete, and are visible to executed SQL commands. trigger <emphasis>will</emphasis> see the effects of data
changes for rows previously processed in the same outer
command. This requires caution, since the ordering of these
change events is not in general predictable; a SQL command that
affects multiple rows may visit the rows in any order.
</para>
</listitem>
<listitem>
<para>
When a row-level after trigger is fired, all data changes made
by the outer command are already complete, and are visible to
the invoked trigger function.
</para> </para>
</listitem> </listitem>
</itemizedlist> </itemizedlist>