Re: pg_stat_statements IN problem

On Tue, 2023-10-03 at 08:05 +0000, byme(at)byme(dot)email wrote:
> "This obfuscates our monitoring because the same query with different amount of arguments gets translated into this:
> IN ($1, $2)
> and so on."
>
> The questions are:
> 1. Shouldnt IN behave so that the query in pg_stat_statements would look like this:
> IN $1
> 2. Shouldnt there be at least some flag to aggregate such queries into one?
> 3. Is there any workaround how to aggregate those queries except the "= ANY"?
> 4. How come no one is bothered by this if this makes pg_stat_statements unusable with lots of queries using IN, what others do with this problem?
> 5. what do you mean by changing pg_stat_statements with another view/table?

There is currently a patch for this very problem under review:

https://commitfest.postgresql.org/44/2837/

The discussion is here:

https://www.postgresql.org/message-id/flat/CA+q6zcWtUbT_Sxj0V6HY6EZ89uv5wuG5aefpe_9n0Jr3VwntFg(at)mail(dot)gmail(dot)com

You could comment on that patch or review it. Useful reviews and supporting
comments help move the patch forward. That would best serve your interests.

Yours,
Laurenz Albe