SQL Server Performance Forum – Threads Archive
Profiler results
I ran the Profiler and now, I have 560,000 row of the following columns: EventClass, TextData, DatabaseID, HostName, ApplicationName, LoginName, SPID, Duration, StartTime, EndTime, Reads, Writes, CPU, ObjectsId, DBUserName I am going to write some SELECT statements with different WHERE clauses or different ORDER BY. How should I analyse the information? For example, my highest Writes is 14175, highest Reads is 3532708 (Which both belong to Import DTS packages). My question is that when these numbers are too much. Does CPU over 5000 means too much or what? Write me your experience and the way you work with such a data. Thanks, CanadaDBAIf I were you, I’ll order by duration, and begin with copy and paste to QA to see execution plan to optimize.
Also you can order by reads, writes, etc., and save as differents files. 560000 rows sound much to me, did you filter by duration? Luis Martin
Moderator
SQL-Server-Performance.com One of the symptoms of an approaching nervous breakdown is the belief that one’s work is terribly important
Bertrand Russell
All postings are provided “AS IS†with no warranties for accuracy.
560000 rows sound much to me, did you filter by duration?
No, I just filtered by CPU >= 1000. And Not like "Profiler" Also you can order by reads, writes, etc., and save as differents files.
The data is in a SQL table. You mean I export them to …? Excel files?
CanadaDBA
1)I suggest to filter by duration.
2)No, I allways capture to file. When finish, I open that file with profiler, look at properties, change the order of the columns (duration by example) and save with other name. Same thing with reads, etc.
In that way I can see what I want with profiler. Luis Martin
Moderator
SQL-Server-Performance.com One of the symptoms of an approaching nervous breakdown is the belief that one’s work is terribly important
Bertrand Russell
All postings are provided “AS IS†with no warranties for accuracy.
I set the Profiler to save data into a file; after the job was finished, I copied the data into a table in SQL-Server on my local machine. I think we did almost the same. Does your approach better than the above? Can I use your approach and filter by duration?
quote:Originally posted by LuisMartin
1)I suggest to filter by duration.
2)No, I allways capture to file. When finish, I open that file with profiler, look at properties, change the order of the columns (duration by example) and save with other name. Same thing with reads, etc.
In that way I can see what I want with profiler. Luis Martin
Moderator
SQL-Server-Performance.com One of the symptoms of an approaching nervous breakdown is the belief that one’s work is terribly important
Bertrand Russell
All postings are provided “AS IS†with no warranties for accuracy.
2)No, I allways capture to file. When finish, I open that file with profiler, look at properties, change the order of the columns (duration by example) and save with other name. Same thing with reads, etc.
In that way I can see what I want with profiler. Luis Martin
Moderator
SQL-Server-Performance.com One of the symptoms of an approaching nervous breakdown is the belief that one’s work is terribly important
Bertrand Russell
All postings are provided “AS IS†with no warranties for accuracy.
CanadaDBA
quote:The data is in a SQL table. You mean I export them to …? Excel files?
I see no need for excel, use all select clauses you mentioned to analyse results. You can save query results in excel if you want to see graphics or use other excel fancy functionality. I never needed that.
When you identify most costly queries you than try to optimize them because their optimization will probably have the most impact.
Yes we did the same. The diference is I don’t use Server resources, I work with files in my PC.
And yes, you can filter using the SQL table you have created.
Luis Martin
Moderator
SQL-Server-Performance.com One of the symptoms of an approaching nervous breakdown is the belief that one’s work is terribly important
Bertrand Russell
All postings are provided “AS IS†with no warranties for accuracy.
My question was that if I use Profiler and read the file (Not SQL table) from my local PC, then can I sort data on the profiler? Can I filter data on the profiler?
quote:Originally posted by LuisMartin
Yes we did the same. The diference is I don’t use Server resources, I work with files in my PC.
And yes, you can filter using the SQL table you have created.
And yes, you can filter using the SQL table you have created.
CanadaDBA
Yes, indeed.
Just open the trace file, properties, go to columns and move column you want to the first place.
Luis Martin
Moderator
SQL-Server-Performance.com One of the symptoms of an approaching nervous breakdown is the belief that one’s work is terribly important
Bertrand Russell
All postings are provided “AS IS†with no warranties for accuracy.
In properties under data columns, I moved the CPU as an example to top of the cloumns. Does it sort the data based on CPU ascendingly? How about if I want to order the data descendingly? CanadaDBA
Only ascendingly, no way to order descendingly.
Luis Martin
Moderator
SQL-Server-Performance.com One of the symptoms of an approaching nervous breakdown is the belief that one’s work is terribly important
Bertrand Russell
All postings are provided “AS IS†with no warranties for accuracy.
What are the measure of Reads, Writes, Duration, and CPU? Are they all mili-seconds? or what? CanadaDBA
Reads, Writes: Number of logical Reads/Writes
Duration, CPU: milisecons.
Luis Martin
Moderator
SQL-Server-Performance.com One of the symptoms of an approaching nervous breakdown is the belief that one’s work is terribly important
Bertrand Russell
All postings are provided “AS IS†with no warranties for accuracy.
Reads, Writes: Number of logical Reads/Writes Logical Reads/Writes? Is it byte or bit or something?
CanadaDBA
It is in Bytes. Satya SKJ
Moderator
http://www.SQL-Server-Performance.Com/forum
This posting is provided “AS IS†with no rights for the sake of knowledge sharing.
As far I know:
Reads: Number of pages read from the data cache.
Writes: Number of pages writes into disk.
Luis Martin
Moderator
SQL-Server-Performance.com One of the symptoms of an approaching nervous breakdown is the belief that one’s work is terribly important
Bertrand Russell
All postings are provided “AS IS†with no warranties for accuracy.
]]>