Hi!
First of all: i'm a real newbie in DB2!

I have a question on measuring the exectuion time of a "select" query. Is there any command to measure it, and, furthermore ,doesn't print out the result?
I'm using DBVisualizer (free edition).

Hope someone knows some help

Martin

go to db2 command window
update monitor switches statement and timestamp
get snapshot for dynamic sql (see infocenter for detail)
will list all executed sql with detailed info
explain will only indicate the cost and not execution time...

Hi,
if using Windows as DB2 client you can use free program DB2 Monitor instead of "get snapshot" command. Download: http://members.tripod.com/chuzhoi_files/index.html
Hope this helps,
Grofaty

Hi Groafty,
thanks for the advice.
Starting the programm I just get the error message, that that DB2APP.dll istn't found (the DB2 distirubtion istn't installed on this computer).

Is there no other easy way to get the execution time of a query?
Sorry, but I don't get this with "get snapshot"...

Regards,
Martin

db2batch is a tool that comes with DB2 to do permformance analysis.

When using db2batch, i have the parameters "Elapsed time" and "CPU Agent time".
Is "CPU Agent Time" the time, which is needed to execute the query without the printout of the resulting tuples?

Regards,
Martin

I would turn off the outputting of rows using the following in the SQL script:
You can also turn off fetch itself with this:
But note that DB2 may actually have to do some work for the fetch if it didn't compute the result set up front. (Computing things up front is not always the best choice, especially if not all rows are retrieved.)

Thanks for the advice!

But which of the two times is then the correct one for me?

Or does there exist still another way to get the pure execution time of a query?

Regards,
Martin

That depends on what you want to know. The "CPU Agent time" may not be correct if your query waits for synchronous I/O, for example.

I would use the "elapsed time" and do away with ROW_OUT to not disturb measurements with writing results to standard output.