Mr. Fuhr,
This appears to be a gremlin. I wasn't able to reproduce the problem. In the interim, I had changed a unique index on the table (took conditional code out of the SQL statement and substituted a function) and wonder if this wasn't interacting with the function. Anyway, it seems to be a false alarm and the good news is that PostgreSQL is back to being its usual, reliable self.
Thanks for the response and sorry to bother.
Michael Schmidt
----- Original Message -----
Sent: Monday, May 22, 2006 11:49 AM
Subject: Re: [GENERAL] Changes in pl/pgsql?
On Mon, May 22, 2006 at 11:20:04AM -0600, Michael Schmidt wrote:
> I've been going through functions for my database, originally
> developed with version 8.0, and had some odd behavior. I'm in the habit
> of putting parentheses around the conditional clauses in IF statements.
> One of my trigger functions didn't work, but when the parentheses were
> deleted, it did. Has something been changed in pl/pgsql for version
> 8.1.3? Here is the statement.
>
> IF (NEW.is_individual) THEN ..
>
> where is_individual is a boolean field
Could you provide a standalone test case and explain what "didn't
work" means? What were you expecting to happen and what actually
did happen? How does 8.1.3's behavior differ from an earlier
version? Exactly which earlier version behaves as expected?
--
Michael Fuhr