I have a MQT which reads data from three tables. The problem is refresh is terribly slow. Note that I have ensured indexes on all the base tables with their statistics updated.

Can anything be done to expedite the refresh?

Thanks in advance

You can take the query SQL of the MQT definition and explain it. Maybe that will help.

Andy

Is that a new measurement unit? And why is that a problem?

You can try temporary disabling logging during refresh table, check this technote for more info:
IBM - How to temporarily turn logging off for operations modifying the table?

But if some error occurs during the not logged operation, the MQT will be marked as inaccessible.

I did an explain on the MQT and it had a paltry cost of 714 timerons but actually if I run the query using db2batch it takes a whopping 723 seconds.

I did try db2advis on the query within the MQT but it did not recommend any indexes as the existing indexes would suffice.

I checked if any of the base tables needed reorganization but it was not the case.

Any inputs on the top of what I have done thus far would be appreciated. Thanks

You could post the SQL for suggestions on how to write it differently. We might be able to help out with something there, but your refresh is going to be as slow or as fast as the query that is being executed can return the rows.
Dave