Re: Duplicate Extended Statistics

Am 04.09.24 um 10:28 schrieb Asad Ali:
> Hi Ilia,
>
> In PostgreSQL, it is possible to create duplicate extended statistics
> because the system does not enforce uniqueness on statistics
> definitions. However, this is generally not recommended, as it leads
> to longer ANALYZE times, increased storage usage, potential planner
> performance impact, and unnecessary complexity. In practice,
> duplicates are rare because users and tools usually avoid redundancy,
> as there is no added benefit to having multiple identical sets of
> statistics on the same columns.
>
> Regards,
> Asad Ali
>
>
To find duplicate indexes:

SELECT  indrelid::regclass::text AS "Table Name",
        string_agg(indexrelid::regclass::text ||
          CASE WHEN indisunique
          THEN ' (UNIQUE)' ELSE ''
          END, ', ') AS "Indexes"
      FROM pg_index
      GROUP BY indrelid, indkey
      HAVING COUNT(*) > 1

--

Holger Jakobs, Bergisch Gladbach