Обсуждение: Query tuning
<div class="Section1"><p class="MsoNormal"><font face="Comic Sans MS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS"">Hi,</span></font><p class="MsoNormal"><font face="Comic Sans MS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS""> </span></font><p class="MsoNormal"><font face="Comic Sans MS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS""> </span></font><p class="MsoNormal"><font face="Comic Sans MS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS"">I have a query which is run across 3 tables JOB_TYPE_FIRST, JOB_TYPE_SECOND and JOB_ALLOCATION_WORKLIST.</span></font><pclass="MsoNormal"><font face="Comic Sans MS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS""> </span></font><p class="MsoNormal"><font face="Comic Sans MS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS"">The column JOBID is referenced in JOB_ALLOCATION_WORKLIST table and primary key in bothJOB_TYPE_FIRST, JOB_TYPE_SECOND tables.</span></font><p class="MsoNormal"><font face="Comic Sans MS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS""> </span></font><p class="MsoNormal"><font face="Comic Sans MS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS"">There is one more column BOOK_ID which is supplied as the binding parameter to the query.The query looks like:</span></font><p class="MsoNormal"><font face="Comic Sans MS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS""> </span></font><p class="MsoNormal"><font face="Comic Sans MS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS"">select count(distinct(j.JOBID)) as jobCount </span></font><p class="MsoNormal"><fontface="Comic Sans MS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS"">from JOB_TYPE_FIRST a, JOB_TYPE_SECOND b, JOB_ALLOCATION_WORKLIST j </span></font><pclass="MsoNormal"><font face="Comic Sans MS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS"">where (( a.JOBID = j.JOBID) </span></font><p class="MsoNormal"><font face="Comic SansMS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS"">and (a.BOOK_ID = :bookId)) </span></font><p class="MsoNormal"><font face="Comic SansMS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS"">or ((b.JOBID = j.JOBID) </span></font><p class="MsoNormal"><font face="Comic Sans MS"size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS"">and (b.BOOK_ID = :bookId));</span></font><p class="MsoNormal"><font face="Comic SansMS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS""> </span></font><p class="MsoNormal"><font face="Comic Sans MS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS"">As the records in the database are too large it is having huge cost and stalling theserver and takes lot of time. </span></font><p class="MsoNormal"><font face="Comic Sans MS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS""> </span></font><p class="MsoNormal"><font face="Comic Sans MS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS"">Can anyone suggest a better way to fetch the results for the query or tune it? Any helpwould be highly appreciated.</span></font><p class="MsoNormal"><font face="Comic Sans MS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS""> </span></font><p class="MsoNormal"><font face="Comic Sans MS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS""> </span></font><p class="MsoNormal"><font face="Comic Sans MS" size="2"><span style="font-size: 10.0pt;font-family:"Comic Sans MS"">Thanks & Regards,</span></font><p class="MsoNormal"><font face="Comic Sans MS" size="2"><spanstyle="font-size: 10.0pt;font-family:"Comic Sans MS"">Kapil</span></font></div><p><strong><span style="font-size:10.0pt;font-family: "Palatino Linotype","serif";color:green"> Please do not print this email unless it is absolutely necessary. </span></strong><spanstyle="font-family:"Arial","sans-serif""></span><p> The information contained in this electronic messageand any attachments to this message are intended for the exclusive use of the addressee(s) and may contain proprietary,confidential or privileged information. If you are not the intended recipient, you should not disseminate, distributeor copy this e-mail. Please notify the sender immediately and destroy all copies of this message and any attachments.<p>WARNING: Computer viruses can be transmitted via email. The recipient should check this email and any attachmentsfor the presence of viruses. The company accepts no liability for any damage caused by any virus transmitted bythis email. <p> www.wipro.com
Here you have specified 3 tables, does JOB_TYPE_FIRST and JOB_TYPE_SECOND both contain all the JOBID in the third table?
Maybe i can help you if you elaborate your problem a bit more.
Regards,
Moiz Kothari
Hi,
I have a query which is run across 3 tables JOB_TYPE_FIRST, JOB_TYPE_SECOND and JOB_ALLOCATION_WORKLIST.
The column JOBID is referenced in JOB_ALLOCATION_WORKLIST table and primary key in both JOB_TYPE_FIRST, JOB_TYPE_SECOND tables.
There is one more column BOOK_ID which is supplied as the binding parameter to the query. The query looks like:
select count(distinct(j.JOBID)) as jobCount
from JOB_TYPE_FIRST a, JOB_TYPE_SECOND b, JOB_ALLOCATION_WORKLIST j
where (( a.JOBID = j.JOBID)
and (a.BOOK_ID = :bookId))
or ((b.JOBID = j.JOBID)
and (b.BOOK_ID = :bookId));
As the records in the database are too large it is having huge cost and stalling the server and takes lot of time.
Can anyone suggest a better way to fetch the results for the query or tune it? Any help would be highly appreciated.
Thanks & Regards,
Kapil
Please do not print this email unless it is absolutely necessary.
The information contained in this electronic message and any attachments to this message are intended for the exclusive use of the addressee(s) and may contain proprietary, confidential or privileged information. If you are not the intended recipient, you should not disseminate, distribute or copy this e-mail. Please notify the sender immediately and destroy all copies of this message and any attachments.
WARNING: Computer viruses can be transmitted via email. The recipient should check this email and any attachments for the presence of viruses. The company accepts no liability for any damage caused by any virus transmitted by this email.
--
Hobby Site : http://dailyhealthtips.blogspot.com
On Thu, May 15, 2008 at 12:00 AM, <kapil.munish@wipro.com> wrote: > Hi, > > select count(distinct(j.JOBID)) as jobCount > from JOB_TYPE_FIRST a, JOB_TYPE_SECOND b, JOB_ALLOCATION_WORKLIST j > where (( a.JOBID = j.JOBID) > and (a.BOOK_ID = :bookId)) > or ((b.JOBID = j.JOBID) > and (b.BOOK_ID = :bookId)); Have you tried joining a with j and b with j separately, and then doing a union on those two sets?
On Fri, 2008-05-16 at 09:21 -0600, Scott Marlowe wrote: > On Thu, May 15, 2008 at 12:00 AM, <kapil.munish@wipro.com> wrote: > > Hi, > > > > select count(distinct(j.JOBID)) as jobCount > > from JOB_TYPE_FIRST a, JOB_TYPE_SECOND b, JOB_ALLOCATION_WORKLIST j > > where (( a.JOBID = j.JOBID) > > and (a.BOOK_ID = :bookId)) > > or ((b.JOBID = j.JOBID) > > and (b.BOOK_ID = :bookId)); > > Have you tried joining a with j and b with j separately, and then > doing a union on those two sets? That seems the only way to get a sane answer to this query, which is otherwise an unconstrained join on both sides of the OR. Great example of a query which runs slow because the question is phrased incorrectly. Count(distinct) is pretty much the only function that will give the same answer as a correctly phrased query. -- Simon Riggs www.2ndQuadrant.comPostgreSQL Training, Services and Support