Re: count(*) and bad design was: Experiences with extensibility

Harald Fuchs wrote:
> In article <60ejcqy6j0.fsf@dba2.int.libertyrms.com>,
> Chris Browne <cbbrowne@acm.org> writes:
>
> > There may be a further optimization to be had by doing a
> > per-statement trigger that counts the number of INSERTs/DELETEs done,
> > so that inserting 30 tuples (in the table being tracked) leads to
> > adding a single tuple with count of 30 in the summary table.
>
> This would be nice, but at least the 8.2.4 docs say
>
>   Statement-level triggers do not currently have any way to examine
>   the individual row(s) modified by the statement.
>
> Is this restriction removed in a later version?

Nope.

--
Alvaro Herrera                                http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.