On 12/9/2011 4:57 PM, David Johnston wrote:
Functions are evaluated once for each row that it generated by the
surrounding query. This is particularly useful if the function in question
takes an aggregate as an input:
SELECT col1, array_processing_function( ARRAY_AGG( col2 ) )
FROM table
GROUP BY col1;
Without this particular behavior you would need to sub-query.
> From a layman's perspective the reason why you cannot use non-aggregates
outside of GROUP BY it that it is ambiguous as to what value to output; with
an uncorrelated function call that is not the case.
David J.
Thanks. This makes sense now. I also went back to the original query
that provoked this question. It had a correlated subquery in the select
statement. I thought that this could yield ambiguous results. But when I
examined it closely, all the correlated fields were included in the
group by of the outer query, and when I tried to use a non-grouped
column from the outer query I correctly got a ERROR: subquery uses
ungrouped column "foo" from outer query
Thanks again.
--
Jack Christensen
jackc@xxxxxxxxxxxxxxxxx
--
Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
