Comments on: Swap values in 2 rows SQL

By: Maresa

Maresa — Wed, 18 Nov 2009 05:15:23 +0000

Yes … I realize that th method doesn’t work when there id unique constraint. I’ve tried many ways and I ended up having to do it in multiple statements.

By: IVO GELOV

IVO GELOV — Mon, 16 Nov 2009 18:54:47 +0000

This method does not work when we have unique constraint/index on the given column

By: Regina

Regina — Thu, 15 Oct 2009 19:50:03 +0000

Small little comment. I think the PostgreSQL create table statement to more mirror the MySQL is

CREATE TABLE rules (
rule_id serial PRIMARY KEY,
rule_name varchar(255) NOT NULL,
priority integer
) ;

the without oids can be left out since its the default behavior for newer versions.

By: Onno Molenkamp

Onno Molenkamp — Mon, 12 Oct 2009 10:25:30 +0000

It can be done with a few where-conditions less:

update rules
set priority = rule1.priority
from rules rule1
where rules.rule_id in (1, 4)
and rule1.rule_id + rules.rule_id = 1 + 4;

By: Maresa

Maresa — Sun, 11 Oct 2009 21:15:47 +0000

The Cartesian product will result a combination of (1, 1), (1,4), (4,1), (4,4). The WHERE condition “rules.rule_id rule1.rule_id” will get you (1, 4) and (4, 1). So the WHERE condition guarantees distinct, unless I’m missing something.

By: Pavel Golub

Pavel Golub — Sun, 11 Oct 2009 20:44:08 +0000

For PostgreSQL you should use DISTINCT FROM I believe… in the last statement of course.