when the first row is ready to return to client, fetch begins and ends until all rows is returned。You need to know that most operation of the executiong plan is executed in a streaming fashion. eg:
select * from t where c1 =xxxx. the plan may looks like this:
at first ,the first page is read from the disk to bufferpool ,then db2 get the first row from the page, and put it to the filter, when it is qulified ,fetch begins。then the second page is read and so on ( we assume there is no data prefetching )
the time after the sql compiled and before the fetch begin is execution time.
Sometimes execution time is very shortly.
if you want to improve your fetch time ,why dont you provide more infomation.
the number of db2batch's fetching time and your application's fetching time.
how do you identify the problem is on fetching time?
how your application fetch the resultset ?
how your application process the record after it has been fetched?
what is the language you used ?? c? java