Re: dynamic table/col names in plpgsql

ugh, I'm going down a rathole with this... the dynamic part of the
query is the table name, and therefore, if I want to select into a row
variable, that variable's declaration needs to be dynamic, too. That
seems kind of crazy, and I see no way to do that anyway. Maybe I'm
going about this all wrong.

So, what is the right approach? The problem I'm trying to solve is
that I want to key off one of the parameters passed into my function,
and then operate on either table_a or table_b, depending on that
parameter (do some select queries, possibly followed by an update
query). The column names are the same for both tables, so all other
parts of the query should remain relatively static. Its just the
table name I need to be dynamic with.

Thanks,
Gary

On 6/25/07, gary jefferson <garyjefferson123(at)gmail(dot)com> wrote:
> Thanks Michael and Steve.
>
> Related question: If I previously had a
>
> IF EXISTS (select ...)
>
> statement, and the 'select ...' part now needs to be run with EXECUTE,
> how do I check for existence? It looks like I need to do an 'INTO'
> with a row variable? I can't seem to find an example snippet of how
> to accomplish this in the pg docs. Where should I be looking?
>
> Thanks again!
> Gary