I want to have a "lastmodified" timestamp (or datetime? not sure if it makes a difference other than presentation of the data) to log the last modified date/time of that record's entry.
Apparently this is possible using triggers. Since I haven't used triggers before, I thought I could first try an "update rule" since that is new to me too:
http://www.postgresql.org/docs/8.3/static/rules-update.html
What I have is this table to log a customer's session data:
CREATE TABLE customer_session (
customer_sessionid serial PRIMARY KEY,
savedsearch_contents text,
lastmodified timestamp default now()
); /*
@ lastmodified - should be updated whenever the table is updated for this entry, just for reference.
*/
Then I could create a rule like this. I'm not sure about the syntax, or whether to use NEW or OLD. Could anyone advise the correct syntax?
CREATE RULE customer_session_lastmodified AS
ON UPDATE TO customer_session
DO UPDATE customer_session SET lastmodified = current_timestamp WHERE customer_sessionid = NEW.customer_sessionid
As you can see I want to update the lastmodified entry of THAT customer_sessionid only, so I'm not sure how to reference it. The UPDATE query would be like this:
UPDATE customer_session SET savedsearch_contents = 'abcde'
WHERE customer_sessionid = {unique customer ID}
Many thanks!
The UPSERT statement is a DBMS feature that allows a DML statement's author to either insert a row or if the row already exists, UPDATE that existing row instead. That is why the action is known as UPSERT (simply a mix of Update and Insert).
In PostgreSQL, the UPDATE command is used to change the present records in a table. To update the selected rows, we have to use the WHERE clause; otherwise, all rows would be updated.
PostgreSQL implements multiversioning by keeping the old version of the table row in the table – an UPDATE adds a new row version (“tuple”) of the row and marks the old version as invalid. In many respects, an UPDATE in PostgreSQL is not much different from a DELETE followed by an INSERT .
You cannot do it with a rule, since it would create an infinite recursion. The correct way is to create a before trigger, just as duffymo proposed.
CREATE FUNCTION sync_lastmod() RETURNS trigger AS $$ BEGIN NEW.lastmodified := NOW(); RETURN NEW; END; $$ LANGUAGE plpgsql; CREATE TRIGGER sync_lastmod BEFORE UPDATE ON customer_session FOR EACH ROW EXECUTE PROCEDURE sync_lastmod();
You could write a trigger that would fire BEFORE UPDATE to modify that date.
See Example 39-4, which adds user name and timestamp before an UPDATE:
http://www.postgresql.org/docs/current/static/plpgsql-trigger.html
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With