short ciruit logic of plpgsql in 8.2

hello,

it depends on SQL. plpgsql hasn't own arithmetic unit. You can test it simply:

postgres=# create or replace function a() returns bool as $$ begin
raise notice 'a'; return true; end$$ language plpgsql;
CREATE FUNCTION
postgres=# create or replace function b() returns bool as $$ begin
raise notice 'a'; return true; end$$ language plpgsql;
CREATE FUNCTION
postgres=# select a() or b();
NOTICE: a
?column?
----------
t
(1 row)

result: statements are short circuited.

2007/5/17, Richard Broersma Jr <rabroersma@yahoo.com>:

Richard Broersma Jr <rabroersma@yahoo.com> writes:

Does anyone know if logic operations in plpgsql are short circuited in 8.2?

They are often optimized, but if you expect "short circuiting" following
exactly the rules of, say, C, you are likely to be disappointed. See
the manual:

http://www.postgresql.org/docs/8.2/static/sql-expressions.html#SYNTAX-EXPRESS-EVAL

regards, tom lane

This keeps biting me. In a trigger function for INSERT OR UPDATE if you try

IF TG_OP = 'UPDATE' AND OLD.foo = 'bar' THEN
...

it will blow up on inserts because there is no OLD. I always expect
this to short circuit and am always disappointed. Easy fix, of
course...

IF TG_OP = 'UPDATE' THEN
IF OLD.foo = 'bar' THEN
...

In case this was what the OP was asking about in particular, it still
happens to me in 8.2

- Ian