Tom Lane <[EMAIL PROTECTED]> writes:
> Robert Creager <[EMAIL PROTECTED]> writes:
> > ... one piece of data I need is the last value for each GROUP BY
> > period. Alas, I cannot figure out how to do this.
>
> SELECT DISTINCT ON (rather than GROUP BY) could get this done for you.
Or if you need to combine this with other aggregate functions like sum, count,
etc:
CREATE FUNCTION first_accum (integer, integer) RETURNS integer AS 'select
coalesce($1,$2)' LANGUAGE sql;
CREATE FUNCTION last_accum (integer, integer) RETURNS integer AS 'select $2' LANGUAGE
sql;
CREATE AGGREGATE first (BASETYPE = integer, SFUNC = first_accum, STYPE = integer);
CREATE AGGREGATE last (BASETYPE = integer, SFUNC = last_accum, STYPE = integer);
Then you can do first() and last(). These definitions only work for integer
but you can pattern match for other datatypes. You might be able to get a
universal function working using anyelement now, I haven't tried.
--
greg
---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match
