Re: Faster "SET search_path"

Robert Haas <robertmhaas@gmail.com>

From: Robert Haas <robertmhaas@gmail.com>

To: Jeff Davis <pgsql@j-davis.com>

Cc: pgsql-hackers@postgresql.org

Date: 2023-07-31T16:55:22Z

Lists: pgsql-hackers

On Sat, Jul 29, 2023 at 11:59 AM Jeff Davis <pgsql@j-davis.com> wrote:
> Unfortunately, adding a "SET search_path" clause to functions slows
> them down. The attached patches close the performance gap
> substantially.

I haven't reviewed the code but I like the concept a lot. This is badly needed.

-- 
Robert Haas
EDB: http://www.enterprisedb.com

Commits

Fix missing invalidations for search_path cache.
- b3bd18294ee4 18.0 landed
- d3e076549b99 17.0 landed
Optimize SearchPathCache by saving the last entry.
- a86c61c9eefa 17.0 landed
Optimize check_search_path() by using SearchPathCache.
- ad57c2a7c586 17.0 landed
Be more paranoid about OOM in search_path cache.
- 8efa301532c8 17.0 landed
Add cache for recomputeNamespacePath().
- f26c2368dcae 17.0 landed
Transform proconfig for faster execution.
- 5765cfe18c59 17.0 landed