Hi, Martin
On Mon, 2021-04-19 at 11:39 +0000, Chengxi Sun wrote:
> I tested your patch, and it works well. I also prefer timestamp
> inseatead of dealloc num.
> I think it can provide more useful details about query statements.
>
Thank you for your review.
Certainly, timestamp is valuable here. Deallocation number is only a
workaround in unlikely case when timestamping will cost a much. It
seems, that it can happen only when significant amount of statements
causes a new entry in pg_stat_statements hashtable. However, in such
case using of pg_stat_statements extension might be qute difficult.
--
Andrei Zubkov
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company