On 2026-Jan-16, Japin Li wrote:

> Thanks for the test. After rethinking, I agree — that BRIN/GIST index-type
> check is useless.
> 
> OTOH, why don't we just do it the same way as the btree index functions?
> 
>       brin_page_items(relname text, blkno bigint);

Operational usefulness.  If you have a corrupt page in a production, you
can extract it from there with get_raw_page() and move it to another
system to do low-level investigation.  If you only have the interface
you suggest, you force the researcher to access the production system
(and run potentially dangerous tools), which may be best avoided.

-- 
Álvaro Herrera         PostgreSQL Developer  —  https://www.EnterpriseDB.com/


Reply via email to