Tom Lane wrote:
Alvaro Herrera <[EMAIL PROTECTED]> writes:
Shane Ambler wrote:
CREATE TABLE jefftest ( id serial, num int );
INSERT INTO jefftest (num) values (generate_series(1,10));
INSERT INTO jefftest (num) values (generate_series(11,20));
INSERT INTO jefftest (num) values (generate_series(21,30));
Don't use set-returning functions in "scalar context".
I think what is actually happening is that the expanded targetlist is
nextval('seq'), generate_series(1,10)
On the eleventh iteration, generate_series() returns ExprEndResult to
show that it's done ... but the 11th nextval() call already happened.
If you switched the columns around, you wouldn't get the extra call.
If you think that's bad, the behavior with multiple set-returning
functions in the same targetlist is even stranger. The whole thing
is a mess and certainly not something we would've invented if we
hadn't inherited it from Berkeley.
regards, tom lane
Would a re-write be something worth adding to the todo list?
and/or maybe add something about this to the manual?
--
Shane Ambler
[EMAIL PROTECTED]
Get Sheeky @ http://Sheeky.Biz
---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match
