Michael Fuhr wrote:
> Why does the first query below return the same value for each row
> while the second query returns random values? Planner optimization?
I assume it is due to some kind of flattening in the planner, but it is
totally unrelated to ARRAY(subquery):
regression=# SELECT (SELECT random()) as f FROM generate_series(1, 5); f
------------------- 0.752416231088534 0.752416231088534 0.752416231088534 0.752416231088534 0.752416231088534
(5 rows)
regression=# SELECT (SELECT random() + 0 * f) as f FROM
generate_series(1, 5) as t(f); f
------------------- 0.176055165555354 0.608546747178094 0.55303416240636 0.127355110425202 0.21671894063089
(5 rows)
Here's another example:
regression=# create table t1(f text);
CREATE TABLE
regression=# insert into t1 values('0');
INSERT 17366 1
regression=# insert into t1 values('0');
INSERT 17367 1
regression=# insert into t1 values('0');
INSERT 17368 1
regression=# SELECT (SELECT timeofday()::timestamp + f::interval) as f
FROM t1; f
---------------------------- 2005-08-16 16:38:12.738215 2005-08-16 16:38:12.738292 2005-08-16 16:38:12.738315
(3 rows)
regression=# SELECT (SELECT timeofday()::timestamp) as f FROM t1; f
---------------------------- 2005-08-16 16:35:33.100791 2005-08-16 16:35:33.100791 2005-08-16 16:35:33.100791
(3 rows)
This shows that it isn't related to use of an SRF either.
Joe
