Re: Disallow redundant indexes

On Mon, Apr 28, 2025, at 7:17 PM, Nathan Bossart wrote:
> On Mon, Apr 28, 2025 at 01:12:18PM -0500, Sami Imseih wrote:
> > If I had a dollar for every time I asked a user to check for duplicate
> > indexes :) So
> > I think it's a good idea to provide some way to detect these ( besides
> > a query in the
> > Wiki ), but I don't think we should prevent it.
> >
> > While a WARNING will be a good to have, it could easily go unnoticed,
> > but it's still good to have.
>
> I don't doubt that this is a problem in the field, but I'm not sure I'm
> convinced it's worthwhile to emit a warning that only appears for exact
> duplicates and is very easily missed, not to mention that the user could be
> intentionally creating the index.

I don't like to idea of a new GUC to control an SQL command. The WARNING seems
fragile because it doesn't solve the problem. I also don't want to disallow
duplicate indexes; it can break old routines. If we want to do something about
it, I would say let's add a psql meta command or even a new view to show
duplicate indexes. I do think that official maintenance queries are a good to
step forward to make life easier for DBAs.

--
Euler Taveira
EDB https://www.enterprisedb.com/