[Info-ingres] A Really Slow Query

Michel Forget mvf at cyberren.com
Fri Oct 19 00:13:24 UTC 2018


Hi Karl,

It took me a minute to find it after setting the trace point since it spits
it out in the middle of the query results, but here it is (though I think
you will have to paste it into notepad for it to look sane, since Outlook
wants this message to be plain text).

Oddly, the query is currently running in what I would consider to be an
acceptable time frame (probably because the data is cached now since we've
run it 90 million times today).  

********************************************************************

                              ----------
                              | 0|
                              HJOIN
                              at     1000
                              et     4549

                              ad     2755
                              ed    48779

                              ac     1906
                              ec  2683549

                              et        1
                              ----------
           /                                      \
          ----------                              ----------
          | 1|                                    | 4|
          HJOIN                                   HJOIN
          at     2614                             at     1000
          et     2610                             et     4549

          ad        0                             ad     2755
          ed       28                             ed    48750

          ac        0                             ac     1918
          ec    10503                             ec  2661426

          et        0                             et        1
          ----------                              ----------
 /                  \                    /                  \
----------          ----------          ----------          ----------
| 2|                | 3|                | 5|                | 6|
ORIG                ORIG                ORIG                HJOIN
at     2615         at     2614         at      627         at     1000
et     2615         et     2614         et      627         et     4549

ad        0         ad        0         ad        0         ad     2755
ed       16         ed       12         ed        2         ed    48748

ac        0         ac        0         ac        0         ac     1920
ec     2615         ec     2614         ec      627         ec  2654936

et        0         et        0         et        0         et        1
----------          ----------          ----------          ----------
                                                   /
\
                                                  ----------
----------
                                                  | 7|
| 8|
                                                  ORIG
TJOIN
                                                  at      627
at     1000
                                                  et      627
et     4549
 

                                                  ad        0
ad     2755
                                                  ed        2
ed    48746
 

                                                  ac        0
ac     1922
                                                  ec      627
ec  2648480
 

                                                  et        0
et        1
                                                  ----------
----------
                                                                       /
 
----------
                                                                      | 9|
                                                                      TSORT
                                                                      at
1037
                                                                      et
45493

                                                                      ad
2755
                                                                      ed
5682

                                                                      ac
1923
                                                                      ec
2557494

                                                                      et
1
 
----------
                                                             /
                                                            ----------
                                                            |10|
                                                            HJOIN
                                                            at     9480
                                                            et    45493

                                                            ad     2755
                                                            ed     5682

                                                            ac     1895
                                                            ec  2557494

                                                            et        1
                                                            ----------
                                                   /                  \
                                                  ----------
----------
                                                  |11|                |12|
                                                  ORIG                ORIG
                                                  at        0         at
9480
                                                  et        1         et
90986

                                                  ad        0         ad
2755
                                                  ed        2         ed
5513

                                                  ac        0         ac
1897
                                                  ec        1         ec
2464120

                                                  et        0         et
1
                                                  ----------
----------

mvf
-----Original Message-----
From: info-ingres-bounces at lists.planetingres.org
<info-ingres-bounces at lists.planetingres.org> On Behalf Of Karl and Betty
Schendel
Sent: Thursday, October 18, 2018 5:51 PM
To: Ingres and related product discussion forum
<info-ingres at lists.planetingres.org>
Subject: Re: [Info-ingres] A Really Slow Query


> On Oct 18, 2018, at 6:55 PM, Michel Forget <mvf at cyberren.com> wrote:
> 
> Hi,
>  
> I'm having a problem with the performance of a query and I'm hoping
someone can help.
>  
> The main problem is that this query takes literally minutes to run.
Looking at the QEP, things seem to go off the rails in the lower right-half
of the QEP.  There is a huge chunk of CPU time on the right-join to the
proj-rest node that got data from secondary index alt2_d4.  There is also
then a huge burst of disk time when it appears Ingres is pulling more data
from base table d4.
>  

What's the QE90 output look like?

Karl

_______________________________________________
Info-ingres mailing list
Info-ingres at lists.planetingres.org
https://lists.planetingres.org/mailman/listinfo/info-ingres




More information about the Info-ingres mailing list