Re: [HACKERS] Thick indexes - a look at count(1) query

2008-01-23 Thread Guillaume Smet
On Jan 23, 2008 5:48 PM, Martijn van Oosterhout <[EMAIL PROTECTED]> wrote: > On Wed, Jan 23, 2008 at 10:02:14PM +0530, Gokulakannan Somasundaram wrote: > > As you said, 'gettimeofday' is occupying 40% of the execution time. But > > without explain analyze(with just executor stats), i am not getting

Re: [HACKERS] Thick indexes - a look at count(1) query

2008-01-23 Thread Martijn van Oosterhout
On Wed, Jan 23, 2008 at 10:02:14PM +0530, Gokulakannan Somasundaram wrote: > As you said, 'gettimeofday' is occupying 40% of the execution time. But > without explain analyze(with just executor stats), i am not getting the time > duration at all. Is there any other way, i can get the timings witho

Re: [HACKERS] Thick indexes - a look at count(1) query

2008-01-23 Thread Gokulakannan Somasundaram
As you said, 'gettimeofday' is occupying 40% of the execution time. But without explain analyze(with just executor stats), i am not getting the time duration at all. Is there any other way, i can get the timings without this overhead? Thanks, Gokul. On Jan 18, 2008 1:23 AM, Decibel! <[EMAIL PROT

Re: [HACKERS] Thick indexes - a look at count(1) query

2008-01-17 Thread Decibel!
On Jan 16, 2008, at 8:07 AM, Gokulakannan Somasundaram wrote: gokul=# explain analyze select count(1) from dd; LOG: EXECUTOR STATISTICS DETAIL: ! system usage stats: ! 0.277208 elapsed 0.275457 user 0.000148 system sec ! [1.128422 user 0.004976 sys total] !

[HACKERS] Thick indexes - a look at count(1) query

2008-01-16 Thread Gokulakannan Somasundaram
Hi, I have submitted a new patch against thick indexes(indexes with snapshot) http://archives.postgresql.org/pgsql-hackers/2007-10/msg00220.php. I did look closely at improving the performance of count(1) queries. It worked well