What are the perfomance gains of using a ProcessFull instead of a
ProcessData and ProcessIndex. It appears they both do the same.

What are the pros and cons of both?




objServer.BeginTransaction()

oNewPartition.Process(ProcessType.ProcessFull)

objServer.CommitTransaction()



vs.





objServer.BeginTransaction()

oNewPartition.Process(ProcessType.ProcessData)

oNewPartition.Process(ProcessType.ProcessIndex)

objServer.CommitTransaction()

In my opinion, not much difference. You may save some time in analyzing and
scheduling the operation if you did a single ProcessFull instead of
splitting it up -- but I doubt that would make much difference.

However, being able to split them up gives you the flexibility to process
them in different transactions -- and what that lets you do is make the new
data available for querying more quickly (albeit with slower query
performance) and build the indexes independently in a separate transaction.

Splitting up the operations may theoretically also make a difference to
processing performance when you are processing multiple partitions in
parallel because different dimension stores would be in memory and perhaps
cause less thrashing and contention -- but this isn't something that is
easily demonstratable or provable although it may be worth experimenting
with...

Thanks,
Akshai
--
Try out the MSDN Forums for Analysis Services at:
http://forums.microsoft.com/MSDN/Sho...ID=83&SiteID=1

This posting is provided "AS IS" with no warranties, and confers no rights
Please do not send email directly to this alias. This alias is for newsgroup
purposes only.

"Jesse O." <jesperzz (AT) hotmail (DOT) com> wrote

Ah gotcha.

I have a cube with three measure groups. It'd make sense to do: ProcessData,
ProcessData, ProcessData, ProcessIndex rather than ProcessFull, ProcessFull,
ProcessFull.

Thanks!


"Akshai Mirchandani [MS]" <akshaim (AT) online (DOT) microsoft.com> wrote

It *may* make sense to do it that way depending on your environment and
needs. In fact, you may perhaps prefer to ProcessData the three measure
groups in parallel in one transaction (using the Batching mechanism) and
then the indexes in parallel in a new transaction. This would typically give
you a shorter processing window but increase load on the system while
processing is going on.

A lot of it depends on what you need to do, and how long your processing
window is...

Thanks,
Akshai

--
Try out the MSDN Forums for Analysis Services at:
http://forums.microsoft.com/MSDN/Sho...ID=83&SiteID=1

This posting is provided "AS IS" with no warranties, and confers no rights
Please do not send email directly to this alias. This alias is for newsgroup
purposes only.

"Jesse O." <jesperzz (AT) hotmail (DOT) com> wrote