On 3/9/26 10:53 AM, Greg Sabino Mullane wrote:
> On Mon, Mar 9, 2026 at 11:37 AM Adrian Klaver <adrian.klaver@aklaver.com
> <mailto:adrian.klaver@aklaver.com>> wrote:
>
> This post in answer:
>
> https://www.postgresql.org/message-id/
> AS2PR05MB107549DDE42DC0B8E31CB52BFEF90A%40AS2PR05MB10754.eurprd05.prod.outlook.com
<https://www.postgresql.org/message-id/AS2PR05MB107549DDE42DC0B8E31CB52BFEF90A%40AS2PR05MB10754.eurprd05.prod.outlook.com>
>
> would seem to indicate that is not the issue.
>
>
> Ah, but those are doing the lookup by the index (via indexrelid) but I
> am curious about looking up by the table (indrelid), as that is how
> pg_dump is going to get at it.
Yeah, but the indrelid did not change after the the REINDEX.
I guess one could question the cast in:
indrelid = 'bcf_work_type'::regclass
Though I am not sure how a REINDEX would affect that?
> Cheers,
> Greg
>
>
--
Adrian Klaver
adrian.klaver@aklaver.com