public inbox for [email protected]
help / color / mirror / Atom feedRe: Slow performance
3+ messages / 3 participants
[nested] [flat]
* Re: Slow performance
@ 2024-07-26 05:30 [email protected] <[email protected]>
2024-07-26 07:54 ` Re: Slow performance Francisco Olarte <[email protected]>
0 siblings, 1 reply; 3+ messages in thread
From: [email protected] @ 2024-07-26 05:30 UTC (permalink / raw)
To: Postgresql General Group <[email protected]>
Hello,Using PG 11.11, One PG Cluster, Windows 2019 Server Standard, Two databases with identical data.
1. First DB: client_db
2. Second DB: client_test
Took backup (pg_dump) of first database (client_db) and restored the database as second database (client_test).
Query:
Select a.examname, a.registrationnumber, b.studentname, d.departmentname, e.levelname, a.subjectcode, c.subjectname, b.regular, a.semester, a.dummynumber, p.semester as curr_sem, a.internalmark, a.externalmark, a.result, coalesce((a.internalmark + a.externalmark),0) as total, a.absent, a.malpractice, c.maxinternalmark, f.noofsemester, a.examstudentstatus, Case When a.result = 'P' Then 'P' Else Case When a.result = 'F' and a.malpractice = 'Y' and a.examstudentstatus is null Then 'M' Else Case When a.result = 'F' and a.absent = 'Y' and a.examstudentstatus = 'R' Then 'R.C' Else Case When a.result = 'F' and a.absent = 'Y' and a.examstudentstatus = 'S' Then 'S.L' Else Case When a.result = 'F' and a.absent = 'Y' and a.examstudentstatus = 'F' Then 'N.P' Else Case When a.result = 'F' and a.absent = 'Y' and a.examstudentstatus = 'W' Then 'W.H' Else 'RA' End End End End End End as res, Concat(RTrim(f.degreeawarded), ' ', RTrim(d.departmentname)) as course, a.revaluation, m.absent as int_abs, n.companyname, n.companydescription, m.totalmark as int_mark, q.addressone, q.addresstwo, Case When a.semester > f.noofsemester Then 'PRIVATE' When a.semester <= f.noofsemester and a.semester in (1,2) Then 'I - Year' When a.semester <= f.noofsemester and a.semester in (3,4) Then 'II - Year' When a.semester <= f.noofsemester and a.semester in (5,6) Then 'III - Year' End as studyr, c.subjectserialno, p.regulararrear From cl_student_exam_subject a Join cl_student_name b On b.companycode = a.companycode And b.registrationnumber = a.registrationnumber Join cl_subject c On c.companycode = a.companycode And c.subjectcode = a.subjectcode Join cl_department_header d On d.departmentheaderpk = b.departmentheaderfk Join cl_level e On e.levelpk = b.Levelfk Join cl_department_detail f On f.departmentheaderfk = b.departmentheaderfk And f.levelfk = b.levelfk Left Outer Join cl_student_internal_mark m On m.companycode = a.companycode And m.registrationnumber = a.registrationnumber And m.subjectcode = a.subjectcode And m.departmentheaderfk = b.departmentheaderfk And m.levelfk = b.levelfk And m.Regular = b.Regular Join co_company n On n.companycode = a.companycode Join cl_student_semester_subject p On p.companycode = a.companycode And p.examheaderfk = a.examheaderfk And p.subjectcode = a.subjectcode And p.registrationnumber = a.registrationnumber And p.semester = a.semester Join co_company_branch q On n.companycode = a.companycode Where a.companycode = '100' And a.examheaderfk = 'BA80952CFF8F4E1C3F9F44B62ED9BF37' And (a.examstudentstatus is null or a.examstudentstatus in ('R', 'S', 'W')) And b.departmentheaderfk in ('04DF8BD89D0844DD4D8AA151EFB28657') And b.levelfk in ('37A9BEC2638844FFD5B1422D83E70EF3') And b.status = 'A' Order By Concat(RTrim(f.degreeawarded), ' ', RTrim(d.departmentname)), Case When a.semester > f.noofsemester Then 'PRIVATE' When a.semester <= f.noofsemester and a.semester in (1,2) Then 'I - Year' When a.semester <= f.noofsemester and a.semester in (3,4) Then 'II - Year' When a.semester <= f.noofsemester and a.semester in (5,6) Then 'III - Year' End, a.registrationnumber, b.regular, p.semester desc, c.subjectserialno, Case When c.subjectcategory = 'T' Then 1 When c.subjectcategory = 'P' Then 2 When c.subjectcategory = 'D' Then 3 When c.subjectcategory = 'V' Then 4 When c.subjectcategory = 'J' Then 5 End, c.ancillary, Case When Substring(a.subjectcode, 6, 1) = 'C' Then 1 When Substring(a.subjectcode, 6, 1) = 'S' Then 2 When Substring(a.subjectcode, 6, 1) = 'A' Then 3 When Substring(a.subjectcode, 6, 1) = 'E' Then 4 Else 5 End, a.subjectcode
Explain Analyze of DB 1 (client_db) :"Sort (cost=2862.35..2862.36 rows=1 width=1088) (actual time=451671.464..451671.495 rows=326 loops=1)"" Sort Key: (concat(rtrim((f.degreeawarded)::text), ' ', rtrim((d.departmentname)::text))), (CASE WHEN (a.semester > f.noofsemester) THEN 'PRIVATE'::text WHEN ((a.semester <= f.noofsemester) AND (a.semester = ANY ('{1,2}'::integer[]))) THEN 'I - Year'::text WHEN ((a.semester <= f.noofsemester) AND (a.semester = ANY ('{3,4}'::integer[]))) THEN 'II - Year'::text WHEN ((a.semester <= f.noofsemester) AND (a.semester = ANY ('{5,6}'::integer[]))) THEN 'III - Year'::text ELSE NULL::text END), a.registrationnumber, b.regular, a.semester DESC, c.subjectserialno, (CASE WHEN (c.subjectcategory = 'T'::bpchar) THEN 1 WHEN (c.subjectcategory = 'P'::bpchar) THEN 2 WHEN (c.subjectcategory = 'D'::bpchar) THEN 3 WHEN (c.subjectcategory = 'V'::bpchar) THEN 4 WHEN (c.subjectcategory = 'J'::bpchar) THEN 5 ELSE NULL::integer END), c.ancillary, (CASE WHEN (""substring""((a.subjectcode)::text, 6, 1) = 'C'::text) THEN 1 WHEN (""substring""((a.subjectcode)::text, 6, 1) = 'S'::text) THEN 2 WHEN (""substring""((a.subjectcode)::text, 6, 1) = 'A'::text) THEN 3 WHEN (""substring""((a.subjectcode)::text, 6, 1) = 'E'::text) THEN 4 ELSE 5 END), a.subjectcode"" Sort Method: quicksort Memory: 193kB"" -> Nested Loop (cost=2.36..2862.34 rows=1 width=1088) (actual time=57829.857..451662.727 rows=326 loops=1)"" -> Nested Loop (cost=2.36..2861.23 rows=1 width=686) (actual time=57829.829..451658.085 rows=326 loops=1)"" Join Filter: (((b.registrationnumber)::text = (p.registrationnumber)::text) AND ((c.subjectcode)::text = (p.subjectcode)::text) AND (a.semester = p.semester))"" Rows Removed by Join Filter: 13614738"" -> Index Scan using ""cl_student_semester_subject_IX3"" on cl_student_semester_subject p (cost=0.55..8.57 rows=1 width=60) (actual time=0.033..55.702 rows=41764 loops=1)"" Index Cond: (((companycode)::text = '100'::text) AND ((examheaderfk)::text = 'BA80952CFF8F4E1C3F9F44B62ED9BF37'::text))"" -> Nested Loop Left Join (cost=1.81..2848.39 rows=244 width=735) (actual time=0.068..10.768 rows=326 loops=41764)"" Join Filter: (((m.departmentheaderfk)::text = (b.departmentheaderfk)::text) AND ((m.levelfk)::text = (b.levelfk)::text) AND (m.regular = b.regular))"" -> Nested Loop (cost=1.39..2135.32 rows=244 width=795) (actual time=0.053..6.723 rows=326 loops=41764)"" -> Nested Loop (cost=1.11..2058.12 rows=244 width=746) (actual time=0.045..4.299 rows=326 loops=41764)"" -> Nested Loop (cost=0.68..312.80 rows=172 width=699) (actual time=0.029..0.338 rows=313 loops=41764)"" -> Seq Scan on co_company n (cost=0.00..1.01 rows=1 width=394) (actual time=0.001..0.001 rows=1 loops=41764)"" Filter: ((companycode)::text = '100'::text)"" -> Nested Loop (cost=0.68..310.07 rows=172 width=305) (actual time=0.025..0.282 rows=313 loops=41764)"" -> Nested Loop (cost=0.27..10.70 rows=1 width=438) (actual time=0.013..0.019 rows=1 loops=41764)"" -> Nested Loop (cost=0.27..9.60 rows=1 width=278) (actual time=0.009..0.013 rows=1 loops=41764)"" -> Seq Scan on cl_department_header d (cost=0.00..1.30 rows=1 width=200) (actual time=0.002..0.004 rows=1 loops=41764)"" Filter: ((departmentheaderpk)::text = '04DF8BD89D0844DD4D8AA151EFB28657'::text)"" Rows Removed by Filter: 23"" -> Index Scan using cl_department_detail_ix1 on cl_department_detail f (cost=0.27..8.29 rows=1 width=78) (actual time=0.007..0.007 rows=1 loops=41764)"" Index Cond: (((departmentheaderfk)::text = '04DF8BD89D0844DD4D8AA151EFB28657'::text) AND ((levelfk)::text = '37A9BEC2638844FFD5B1422D83E70EF3'::text))"" -> Seq Scan on cl_level e (cost=0.00..1.09 rows=1 width=160) (actual time=0.001..0.003 rows=1 loops=41764)"" Filter: ((levelpk)::text = '37A9BEC2638844FFD5B1422D83E70EF3'::text)"" Rows Removed by Filter: 6"" -> Index Scan using cl_student_name_ix4 on cl_student_name b (cost=0.41..297.65 rows=172 width=97) (actual time=0.011..0.165 rows=313 loops=41764)"" Index Cond: (((companycode)::text = '100'::text) AND ((departmentheaderfk)::text = '04DF8BD89D0844DD4D8AA151EFB28657'::text) AND ((levelfk)::text = '37A9BEC2638844FFD5B1422D83E70EF3'::text) AND (status = 'A'::bpchar))"" -> Index Scan using ""cl_student_exam_subject_IX1"" on cl_student_exam_subject a (cost=0.42..10.13 rows=2 width=89) (actual time=0.012..0.012 rows=1 loops=13072132)"" Index Cond: (((companycode)::text = '100'::text) AND ((examheaderfk)::text = 'BA80952CFF8F4E1C3F9F44B62ED9BF37'::text) AND ((registrationnumber)::text = (b.registrationnumber)::text))"" Filter: ((examstudentstatus IS NULL) OR (examstudentstatus = ANY ('{R,S,W}'::bpchar[])))"" -> Index Scan using cl_subject_ix3 on cl_subject c (cost=0.28..0.32 rows=1 width=53) (actual time=0.007..0.007 rows=1 loops=13615064)"" Index Cond: (((companycode)::text = '100'::text) AND ((subjectcode)::text = (a.subjectcode)::text))"" -> Index Scan using ""cl_student_internal_mark_IX"" on cl_student_internal_mark m (cost=0.42..2.90 rows=1 width=97) (actual time=0.011..0.011 rows=1 loops=13615064)"" Index Cond: (((companycode)::text = (a.companycode)::text) AND ((companycode)::text = '100'::text) AND ((subjectcode)::text = (a.subjectcode)::text) AND ((registrationnumber)::text = (a.registrationnumber)::text))"" Filter: (((departmentheaderfk)::text = '04DF8BD89D0844DD4D8AA151EFB28657'::text) AND ((levelfk)::text = '37A9BEC2638844FFD5B1422D83E70EF3'::text))"" -> Seq Scan on co_company_branch q (cost=0.00..1.01 rows=1 width=276) (actual time=0.001..0.001 rows=1 loops=326)""Planning Time: 15.936 ms""Execution Time: 451672.059 ms"
Explain Analyze of Second DB (client_test)"Sort (cost=3454.91..3454.92 rows=1 width=1088) (actual time=19.120..19.137 rows=326 loops=1)"" Sort Key: (concat(rtrim((f.degreeawarded)::text), ' ', rtrim((d.departmentname)::text))), (CASE WHEN (a.semester > f.noofsemester) THEN 'PRIVATE'::text WHEN ((a.semester <= f.noofsemester) AND (a.semester = ANY ('{1,2}'::integer[]))) THEN 'I - Year'::text WHEN ((a.semester <= f.noofsemester) AND (a.semester = ANY ('{3,4}'::integer[]))) THEN 'II - Year'::text WHEN ((a.semester <= f.noofsemester) AND (a.semester = ANY ('{5,6}'::integer[]))) THEN 'III - Year'::text ELSE NULL::text END), a.registrationnumber, b.regular, a.semester DESC, c.subjectserialno, (CASE WHEN (c.subjectcategory = 'T'::bpchar) THEN 1 WHEN (c.subjectcategory = 'P'::bpchar) THEN 2 WHEN (c.subjectcategory = 'D'::bpchar) THEN 3 WHEN (c.subjectcategory = 'V'::bpchar) THEN 4 WHEN (c.subjectcategory = 'J'::bpchar) THEN 5 ELSE NULL::integer END), c.ancillary, (CASE WHEN (""substring""((a.subjectcode)::text, 6, 1) = 'C'::text) THEN 1 WHEN (""substring""((a.subjectcode)::text, 6, 1) = 'S'::text) THEN 2 WHEN (""substring""((a.subjectcode)::text, 6, 1) = 'A'::text) THEN 3 WHEN (""substring""((a.subjectcode)::text, 6, 1) = 'E'::text) THEN 4 ELSE 5 END), a.subjectcode"" Sort Method: quicksort Memory: 193kB"" -> Nested Loop (cost=2.23..3454.90 rows=1 width=1088) (actual time=0.319..14.984 rows=326 loops=1)"" -> Nested Loop (cost=2.23..3453.78 rows=1 width=686) (actual time=0.298..13.691 rows=326 loops=1)"" Join Filter: (((b.registrationnumber)::text = (p.registrationnumber)::text) AND ((c.subjectcode)::text = (p.subjectcode)::text))"" -> Nested Loop Left Join (cost=1.81..2739.81 rows=230 width=735) (actual time=0.259..9.970 rows=326 loops=1)"" Join Filter: (((m.departmentheaderfk)::text = (b.departmentheaderfk)::text) AND ((m.levelfk)::text = (b.levelfk)::text) AND (m.regular = b.regular))"" -> Nested Loop (cost=1.39..2060.47 rows=230 width=795) (actual time=0.233..6.232 rows=326 loops=1)"" -> Nested Loop (cost=1.11..1987.62 rows=230 width=746) (actual time=0.211..3.955 rows=326 loops=1)"" -> Nested Loop (cost=0.68..299.78 rows=166 width=699) (actual time=0.118..0.393 rows=313 loops=1)"" -> Seq Scan on co_company n (cost=0.00..1.01 rows=1 width=394) (actual time=0.026..0.027 rows=1 loops=1)"" Filter: ((companycode)::text = '100'::text)"" -> Nested Loop (cost=0.68..297.11 rows=166 width=305) (actual time=0.091..0.316 rows=313 loops=1)"" -> Nested Loop (cost=0.27..10.70 rows=1 width=438) (actual time=0.041..0.048 rows=1 loops=1)"" -> Nested Loop (cost=0.27..9.60 rows=1 width=278) (actual time=0.034..0.038 rows=1 loops=1)"" -> Seq Scan on cl_department_header d (cost=0.00..1.30 rows=1 width=200) (actual time=0.009..0.011 rows=1 loops=1)"" Filter: ((departmentheaderpk)::text = '04DF8BD89D0844DD4D8AA151EFB28657'::text)"" Rows Removed by Filter: 23"" -> Index Scan using cl_department_detail_ix1 on cl_department_detail f (cost=0.27..8.29 rows=1 width=78) (actual time=0.024..0.024 rows=1 loops=1)"" Index Cond: (((departmentheaderfk)::text = '04DF8BD89D0844DD4D8AA151EFB28657'::text) AND ((levelfk)::text = '37A9BEC2638844FFD5B1422D83E70EF3'::text))"" -> Seq Scan on cl_level e (cost=0.00..1.09 rows=1 width=160) (actual time=0.006..0.008 rows=1 loops=1)"" Filter: ((levelpk)::text = '37A9BEC2638844FFD5B1422D83E70EF3'::text)"" Rows Removed by Filter: 6"" -> Index Scan using cl_student_name_ix4 on cl_student_name b (cost=0.41..284.75 rows=166 width=97) (actual time=0.049..0.184 rows=313 loops=1)"" Index Cond: (((companycode)::text = '100'::text) AND ((departmentheaderfk)::text = '04DF8BD89D0844DD4D8AA151EFB28657'::text) AND ((levelfk)::text = '37A9BEC2638844FFD5B1422D83E70EF3'::text) AND (status = 'A'::bpchar))"" -> Index Scan using ""cl_student_exam_subject_IX1"" on cl_student_exam_subject a (cost=0.42..10.15 rows=2 width=89) (actual time=0.010..0.011 rows=1 loops=313)"" Index Cond: (((companycode)::text = '100'::text) AND ((examheaderfk)::text = 'BA80952CFF8F4E1C3F9F44B62ED9BF37'::text) AND ((registrationnumber)::text = (b.registrationnumber)::text))"" Filter: ((examstudentstatus IS NULL) OR (examstudentstatus = ANY ('{R,S,W}'::bpchar[])))"" -> Index Scan using cl_subject_ix3 on cl_subject c (cost=0.28..0.32 rows=1 width=53) (actual time=0.006..0.006 rows=1 loops=326)"" Index Cond: (((companycode)::text = '100'::text) AND ((subjectcode)::text = (a.subjectcode)::text))"" -> Index Scan using ""cl_student_internal_mark_IX"" on cl_student_internal_mark m (cost=0.42..2.94 rows=1 width=97) (actual time=0.010..0.010 rows=1 loops=326)"" Index Cond: (((companycode)::text = (a.companycode)::text) AND ((companycode)::text = '100'::text) AND ((subjectcode)::text = (a.subjectcode)::text) AND ((registrationnumber)::text = (a.registrationnumber)::text))"" Filter: (((departmentheaderfk)::text = '04DF8BD89D0844DD4D8AA151EFB28657'::text) AND ((levelfk)::text = '37A9BEC2638844FFD5B1422D83E70EF3'::text))"" -> Index Scan using ""cl_student_semester_subject_IX1"" on cl_student_semester_subject p (cost=0.42..3.09 rows=1 width=60) (actual time=0.010..0.010 rows=1 loops=326)"" Index Cond: (((companycode)::text = '100'::text) AND ((subjectcode)::text = (a.subjectcode)::text) AND ((registrationnumber)::text = (a.registrationnumber)::text) AND (semester = a.semester))"" Filter: ((examheaderfk)::text = 'BA80952CFF8F4E1C3F9F44B62ED9BF37'::text)"" -> Seq Scan on co_company_branch q (cost=0.00..1.01 rows=1 width=276) (actual time=0.000..0.000 rows=1 loops=326)""Planning Time: 15.581 ms""Execution Time: 19.546 ms"
The query when run against DB1 takes around 7 min 32 seconds.The same query when run against DB2 takes around 124 msec.
Same computer, same PG cluster, same query.Why it takes so much time when run against DB1 (client_db)?
Already executed vacuum against client_db database.
Any help is really appreciated.
Happiness Always
BKR Sivaprakash
^ permalink raw reply [nested|flat] 3+ messages in thread
* Re: Slow performance
2024-07-26 05:30 Re: Slow performance [email protected] <[email protected]>
@ 2024-07-26 07:54 ` Francisco Olarte <[email protected]>
2024-07-26 09:40 ` Re: Slow performance Fatih Sazan <[email protected]>
0 siblings, 1 reply; 3+ messages in thread
From: Francisco Olarte @ 2024-07-26 07:54 UTC (permalink / raw)
To: [email protected] <[email protected]>; +Cc: Postgresql General Group <[email protected]>
Hello:
On Fri, 26 Jul 2024 at 07:31, [email protected]
<[email protected]> wrote:
...
> Took backup (pg_dump) of first database (client_db) and restored the database as second database (client_test).
...
> The query when run against DB1 takes around 7 min 32 seconds.
> The same query when run against DB2 takes around 124 msec.
> Same computer, same PG cluster, same query.
> Why it takes so much time when run against DB1 (client_db)?
Can be bad luck, but the usual suspect would be different databases.
I assume db1 is quiescent on the tests ( as it seems the production
database, no heavy querying concurrent with your tests ).
Bear in mind restoring leaves the database similar to what a vacuum
full will do, so it can differ a lot from the original.
> Already executed vacuum against client_db database.
I think you already have pointed out this, but IIRC you have not told
us if you have ANALYZED any of the databases. This is important. Bad
stats in any of them could make the planner choose a bad plan ( or, if
you are unlucky, make it choose a bad one ).
Also, did you vacuum verbose? where your tables well packed? ( bad
vacuuming can lead to huge tables with a lot of free space, but I
doubt this is your case, but everything has to be checked, we only
know what you write us ).
And now, not being an expert in tracing explain I see this in plan-db1:
" Join Filter: (((b.registrationnumber)::text =
(p.registrationnumber)::text) AND ((c.subjectcode)::text =
(p.subjectcode)::text) AND (a.semester = p.semester))"
" Rows Removed by Join Filter: 13614738"
" -> Index Scan using
""cl_student_semester_subject_IX3"" on cl_student_semester_subject p
(cost=0.55..8.57 rows=1 width=60) (actual time=0.033..55.702
rows=41764 loops=1)"
" Index Cond: (((companycode)::text = '100'::text)
AND ((examheaderfk)::text =
'BA80952CFF8F4E1C3F9F44B62ED9BF37'::text))"
Not an explain expert, but if i read correctly an index scan expecting
1 row recovers 41674, which hints at bad statistics ( or skewed data
distribution and bad luck )
The plans are similar, but in the fast query
cl_student_semester_subject is accessed using other index:
" -> Index Scan using
""cl_student_semester_subject_IX1"" on cl_student_semester_subject p
(cost=0.42..3.09 rows=1 width=60) (actual time=0.010..0.010 rows=1
loops=326)"
" Index Cond: (((companycode)::text = '100'::text)
AND ((subjectcode)::text = (a.subjectcode)::text) AND
((registrationnumber)::text = (a.registrationnumber)::text) AND
(semester = a.semester))"
Which seems much more selective and recovers just what it wants.
I would start by analyzing ( and, if not too costly, reindexing ) that table.
Francisco Olarte.
^ permalink raw reply [nested|flat] 3+ messages in thread
* Re: Slow performance
2024-07-26 05:30 Re: Slow performance [email protected] <[email protected]>
2024-07-26 07:54 ` Re: Slow performance Francisco Olarte <[email protected]>
@ 2024-07-26 09:40 ` Fatih Sazan <[email protected]>
0 siblings, 0 replies; 3+ messages in thread
From: Fatih Sazan @ 2024-07-26 09:40 UTC (permalink / raw)
To: [email protected] <[email protected]>; +Cc: Postgresql General Group <[email protected]>
Hi Siva,
pg_dump taken from client_db seems not to have transferred full data to
client_test.
When I examine the query plans, the rows scanned seem to be extremely
different. For example, there is 1 row in cl_level table on client_test,
while clined_db shows that around 300,000 records were scanned.
My suggestion would be to check the data counts in the tables you moved
with count(*).
Francisco Olarte <[email protected]>, 26 Tem 2024 Cum, 10:55 tarihinde
şunu yazdı:
> Hello:
>
> On Fri, 26 Jul 2024 at 07:31, [email protected]
> <[email protected]> wrote:
> ...
> > Took backup (pg_dump) of first database (client_db) and restored the
> database as second database (client_test).
> ...
> > The query when run against DB1 takes around 7 min 32 seconds.
> > The same query when run against DB2 takes around 124 msec.
> > Same computer, same PG cluster, same query.
> > Why it takes so much time when run against DB1 (client_db)?
>
> Can be bad luck, but the usual suspect would be different databases.
>
> I assume db1 is quiescent on the tests ( as it seems the production
> database, no heavy querying concurrent with your tests ).
>
> Bear in mind restoring leaves the database similar to what a vacuum
> full will do, so it can differ a lot from the original.
>
> > Already executed vacuum against client_db database.
>
> I think you already have pointed out this, but IIRC you have not told
> us if you have ANALYZED any of the databases. This is important. Bad
> stats in any of them could make the planner choose a bad plan ( or, if
> you are unlucky, make it choose a bad one ).
>
> Also, did you vacuum verbose? where your tables well packed? ( bad
> vacuuming can lead to huge tables with a lot of free space, but I
> doubt this is your case, but everything has to be checked, we only
> know what you write us ).
>
> And now, not being an expert in tracing explain I see this in plan-db1:
> " Join Filter: (((b.registrationnumber)::text =
> (p.registrationnumber)::text) AND ((c.subjectcode)::text =
> (p.subjectcode)::text) AND (a.semester = p.semester))"
> " Rows Removed by Join Filter: 13614738"
> " -> Index Scan using
> ""cl_student_semester_subject_IX3"" on cl_student_semester_subject p
> (cost=0.55..8.57 rows=1 width=60) (actual time=0.033..55.702
> rows=41764 loops=1)"
> " Index Cond: (((companycode)::text = '100'::text)
> AND ((examheaderfk)::text =
> 'BA80952CFF8F4E1C3F9F44B62ED9BF37'::text))"
>
> Not an explain expert, but if i read correctly an index scan expecting
> 1 row recovers 41674, which hints at bad statistics ( or skewed data
> distribution and bad luck )
>
> The plans are similar, but in the fast query
> cl_student_semester_subject is accessed using other index:
>
> " -> Index Scan using
> ""cl_student_semester_subject_IX1"" on cl_student_semester_subject p
> (cost=0.42..3.09 rows=1 width=60) (actual time=0.010..0.010 rows=1
> loops=326)"
> " Index Cond: (((companycode)::text = '100'::text)
> AND ((subjectcode)::text = (a.subjectcode)::text) AND
> ((registrationnumber)::text = (a.registrationnumber)::text) AND
> (semester = a.semester))"
>
> Which seems much more selective and recovers just what it wants.
>
> I would start by analyzing ( and, if not too costly, reindexing ) that
> table.
>
> Francisco Olarte.
>
>
>
^ permalink raw reply [nested|flat] 3+ messages in thread
end of thread, other threads:[~2024-07-26 09:40 UTC | newest]
Thread overview: 3+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2024-07-26 05:30 Re: Slow performance [email protected] <[email protected]>
2024-07-26 07:54 ` Francisco Olarte <[email protected]>
2024-07-26 09:40 ` Fatih Sazan <[email protected]>
This inbox is served by agora; see mirroring instructions
for how to clone and mirror all data and code used for this inbox