Re: "Group By " index usage

sdfasdfas sdfasdfs <descuarinjado(at)yahoo(dot)es> writes:
> I have a table with this index:
> create index ARTISTS_NAME on ARTISTS (
> lower(AR_NAME)
> );

> Te index is over a colum with this definition:

> AR_NAME VARCHAR(256) null,

> I want to optimize this query:

> select * from artists where lower(ar_name) like
> lower('a%') order by lower(ar_name) limit 20;

> I think the planner should use the index i have.

Update to 7.4, or declare the column as TEXT instead of VARCHAR.
Older versions aren't very bright about situations involving
implicit coercions.

regards, tom lane