CREATE RULE name AS ON event TO object [ WHERE condition ] DO [ INSTEAD ] action where action can be: NOTHING | query | ( query ; query ... ) | [ query ; query ... ]
The name of a rule to create.
Event is one of SELECT, UPDATE, DELETE or INSERT.
Object is either table or table.column. (Currently, only the table form is actually implemented.)
Any SQL boolean-condition expression. The condition expression may not refer to any tables except new and old.
The query or queries making up the action can be any SQL SELECT, INSERT, UPDATE, DELETE, or NOTIFY statement.
Within the condition and action, the special table names new and old may be used to refer to values in the referenced table (the object). new is valid in ON INSERT and ON UPDATE rules to refer to the new row being inserted or updated. old is valid in ON SELECT, ON UPDATE, and ON DELETE rules to refer to the existing row being selected, updated, or deleted.
The Postgres rule system allows one to define an alternate action to be performed on inserts, updates, or deletions from database tables. Rules are used to implement table views as well.
The semantics of a rule is that at the time an individual instance (row) is accessed, inserted, updated, or deleted, there is an old instance (for selects, updates and deletes) and a new instance (for inserts and updates). All the rules for the given event type and the given target object (table) are examined, in an unspecified order. If the condition specified in the WHERE clause (if any) is true, the action part of the rule is executed. The action is done instead of the original query if INSTEAD is specified; otherwise it is done before the original query is performed. Within both the condition and action, values from fields in the old instance and/or the new instance are substituted for old.attribute-name and new.attribute-name.
The action part of the rule can consist of one or more queries. To write multiple queries, surround them with either parentheses or square brackets. Such queries will be performed in the specified order (whereas there are no guarantees about the execution order of multiple rules for an object). The action can also be NOTHING indicating no action. Thus, a DO INSTEAD NOTHING rule suppresses the original query from executing (when its condition is true); a DO NOTHING rule is useless.
The action part of the rule executes with the same command and transaction identifier as the user command that caused activation.
Presently, ON SELECT rules must be unconditional INSTEAD rules and must have actions that consist of a single SELECT query. Thus, an ON SELECT rule effectively turns the object table into a view, whose visible contents are the rows returned by the rule's SELECT query rather than whatever had been stored in the table (if anything). It is considered better style to write a CREATE VIEW command than to create a table and define an ON SELECT rule for it.
You must have rule definition access to a table in order to define a rule on it. Use GRANT and REVOKE to change permissions.
It is very important to take care to avoid circular rules. For example, though each of the following two rule definitions are accepted by Postgres, the select command will cause Postgres to report an error because the query cycled too many times:
Example 1. Example of a circular rewrite rule combination:
CREATE RULE bad_rule_combination_1 AS ON SELECT TO emp DO INSTEAD SELECT * FROM toyemp;
CREATE RULE bad_rule_combination_2 AS ON SELECT TO toyemp DO INSTEAD SELECT * FROM emp;
This attempt to select from EMP will cause Postgres to issue an error because the queries cycled too many times:
SELECT * FROM emp;
If you see anything in the documentation that is not correct, does not match your experience with the particular feature or requires further clarification, please use this form to report a documentation issue.