I was wondering if dropping a table is more efficient in PostgreSQL
8.x in comparison to deleting it's content ?

To put you in the picture - I want to schedule via cron job some daily
data materialization. With the resource I got there are only those two
ways:

1) DROP statement:
a) DROP TABLE tmp;
b) CREATE TABLE tmp AS select * from ....;

2) DELETE statement:
a) DELETE FROM tmp;
b) INSERT INTO tmp SELECT * from ....;

tmp table doesn't have any index nor constraints.

The thing is, postgresql may leave some invalid content behind in both
situations. The real question is - which of those two options leaves
less garbage to be vaccumed ? At this point I don't relay care about
cost based efficiency but cutting down pg background work.

Responses

• Re: Yet another "drop table vs delete" question at 2009-04-21 19:28:28 from Peter Eisentraut
• Re: Yet another "drop table vs delete" question at 2009-04-21 19:30:06 from Alvaro Herrera

pgsql-general by date

Next:	From: Robert Morton	Date: 2009-04-21 17:32:36
	Subject: Re: round behavior differs between 8.1.5 and 8.3.7
Previous:	From: Alvaro Herrera	Date: 2009-04-21 17:13:38
	Subject: Re: trouble with to_char('L')