Tom Lane wrote:
Peter <[EMAIL PROTECTED]> writes:
2. No. In fact I can replicate the problem on the same computer using
PgAdmin (connecting as two different users)
3. I can isolate the part but it does not help much. I'm even doing
RAISE NOTICE and printing out the actual SQL statement that is being
executed... and it's IDENTICAL for both users yet returns different
values. Executing exactly the same SQL statement from commandline
produces identical results for both users.
It's really pretty hard to believe that the same function would work
differently for two different users. I'm wondering if maybe the two
users aren't executing the same function --- perhaps not the specific
one you are looking at, but some other one it depends on. Do the
users have different search_path settings?
I'm thinking that the data returned might be dependent upon the user. Is
this the case? And can any of those columns used in the proc be NULL?
Frankly, i couldn't read it very well (but then, i'm on my first cup of
coffee this morning), but i'm wondering if the problem lies somewhere in
between the dollar-quoting and coalesce().
brian
---------------------------(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
