> Did you run the select remotely on a Windows server?

> Yes the server load is practically 0. Note the difference between
> and
> remote execution of the command. I think you are right about the
> problem possibility. But it is bound to PostgreSQL. MySQL on the same
> machine (and same database content) had no problem.
> So are there any known issues with PostgreSQL on Windows sending data
> remote hosts connected via ODBC?
> What should I do to find out more debug info?

1. turn on all your logging and make sure we looking at the right place
(planner stats, etc).
2. run explain analyze and compare timings (which returns only explain
3. do a select max(patternxml) test.t_umlpattern and observe the time.
4. do a select substr(patternxml, 1, 10) from test.t_umlpattern and
observe the time.
5. do select array_accum(q::text) from generate_series(1,10000) q;

if array_accum errors out, do:

CREATE AGGREGATE public.array_accum (
    sfunc = array_append,
    basetype = anyelement,
    stype = anyarray,
    initcond = '{}'

and observe the time.


---------------------------(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

Reply via email to