Re: REINDEX filtering in the backend - Mailing list pgsql-hackers

On Wed, Aug 28, 2019 at 02:02:08PM +0900, Michael Paquier wrote:
> +       index = index_open(indexOid, AccessShareLock);
> +       numAtts = index->rd_index->indnatts;
> +       index_close(index, AccessShareLock);
> Wouldn't it be better to close that after doing the scan?
>
> Nit: I am pretty sure that this should be indented.
>
> Could you add tests with REINDEX CONCURRENTLY?

Bonus: support for reindexdb should be added.  Let's not forget about
it.
--
Michael