Re: Queries for unused/useless indexes

On 2015-05-25 12:25:01 -0400, Melvin Davidson wrote:
> I'm not sure why you are using "pg_stat_user_indexes".

Because you did. I didn't change that.

> My original query below
> uses "pg_stat_all_indexes" and the schema names are joined and it does work.

I'm not sure what you mean by "original", but this:

> SELECT n.nspname as schema,
>        i.relname as table,
>        i.indexrelname as index,
>        i.idx_scan,
>        i.idx_tup_read,
>        i.idx_tup_fetch,
>        pg_size_pretty(pg_relation_size(quote_ident(n.nspname) || '.' ||
> quote_ident(i.relname))) AS table_size,
>        pg_size_pretty(pg_relation_size(quote_ident(n.nspname) || '.' ||
> quote_ident(i.indexrelname))) AS index_size,
>        pg_get_indexdef(idx.indexrelid) as idx_definition
>   FROM pg_stat_all_indexes i
>   JOIN pg_class c ON (c.oid = i.relid)
>   JOIN pg_namespace n ON (n.oid = c.relnamespace)
>   JOIN pg_index idx ON (idx.indexrelid =  i.indexrelid )
>  WHERE i.idx_scan < 200
>    AND NOT idx.indisprimary
>    AND NOT idx.indisunique
>  ORDER BY 1, 2, 3;

is not the query you posted in your original message.

Here is what you posted:

> On Mon, May 25, 2015 at 10:41 AM, Peter J. Holzer <hjp-pgsql(at)hjp(dot)at> wrote:
>
> On 2015-05-22 09:41:57 -0400, Melvin Davidson wrote:
> > I'd like to share those queries with the community, as I know there must
> be
> > others out there with the same problem.
> >
> > /* useless_indexes.sql */
> > SELECT
> >        idstat.schemaname AS schema,
> >        idstat.relname AS table_name,
> >        indexrelname AS index_name,
> >        idstat.idx_scan AS times_used,
> >        pg_size_pretty(pg_relation_size(quote_ident(idstat.schemaname) ||
> '.' ||
> > quote_ident(idstat.relname))) AS table_size,
> >        pg_size_pretty(pg_relation_size(quote_ident(idstat.schemaname) ||
> '.' ||
> > quote_ident(indexrelname))) AS index_size,
> >        n_tup_upd + n_tup_ins + n_tup_del as num_writes,
> >        indexdef AS definition
> > FROM pg_stat_user_indexes AS idstat
> > JOIN pg_indexes ON indexrelname = indexname
> > JOIN pg_stat_user_tables AS tabstat ON idstat.relname = tabstat.relname
> > WHERE idstat.idx_scan < 200
> > AND indexdef !~* 'unique'
> > ORDER BY idstat.schemaname,
> >          idstat.relname,
> >          indexrelname;

--
_ | Peter J. Holzer | I want to forget all about both belts and
|_|_) | | suspenders; instead, I want to buy pants
| | | hjp(at)hjp(dot)at | that actually fit.
__/ | http://www.hjp.at/ | -- http://noncombatant.org/