andi@kretzer-berlin.de wrote:
>...
> There seems to be no way to find out if there are just no matching rows
> ...
It's easier with pg_execute, since it returns the number of tuples. This
new command showed up around PostgreSQL 7.1, and it's still there at 7.2,
but still undocumented. It looks rather useful. Another thing it can do,
that you can't otherwise do with libpgtcl, is find out how many rows were
affected by insert, update, or delete. I can see pg_execute replacing
pg_select because it is more flexible, too.
Since it isn't listed in HISTORY or the documentation, I wonder if it is
considered "experimental", or might it be staying around, just waiting for
somebody to document it? I would like to know, because there is a bug in
the "-oid" code; I have a tiny patch for it if it is going to be supported.