But if you do build an index over "id" then pgsql would only have to do a
sequential scan on that index, which might be a lot faster if your table
contains a lot of other data, won't it?
Jerry
""Ed L."" <pgsql@bluepolka.net> wrote in message
news:200411060930.30859.pgsql@bluepolka.net...
> On Saturday November 6 2004 7:34, Net Virtual Mailing Lists wrote:
>> Is there a way to create an index that would make this query be efficient
>> and not perform a sequential scan?
>>
>> SELECT count(*) AS count,id FROM sometable GROUP BY id;
>>
>> .. I've considered creating a rule on this table which would put the
>> results of this into another table anytime it is updated, but I thought
>> there might be an easier way.
>
> Since you have no "WHERE" clause and you want to group by id, I believe
> pgsql has to scan all id values. Those id values are only fully stored in
> the table, so I don't think so.
>
> Ed
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 7: don't forget to increase your free space map settings
>
