Hi,

I have to process a very large dimension (~200,000,000 rows). ProcessFull for this dimension take 6 hours

I can identified new members, so I do a ProcessAdd every day. But, ProcessAdd take longer than ProcessFull.

I dont't want to do a ProcessFull every day because I don't want to re-process all my partition in cube (~2,000,000,000 rows).

How can I optimize ProcessAdd ? Why it's so long to process ?

thanks

Hi S,

I suspect that the ProcessFull is faster because it simply trashes the old dimension and invalidates the connected measure groups. Your ProcessAdd, on the other hand, builds the new shadow dimension and "connects" it to your existing measure groups (within a transaction) and leaves the cube available to read. As you might imagine, this requires some extra work.

Hope that helps,

Hi,

Process Add-updates dimension attribute captions and descriptions. update  process is slow.

Process Full-Analysis Services drops all data in the object, and then processes the object

Ref-https://msdn.microsoft.com/en-IN/library/ms174774.aspx

as we know that insert is faster than update.

Regards,

Manish

Hi,

Please see also- in Case of Process update and process full also happen same

https://social.msdn.microsoft.com/Forums/sqlserver/en-US/fd95184c-6e13-487a-a908-5da9205697ff/processupdate-compared-to-processfull-on-single-partition?forum=sqlanalysisservices

Regards,

Manish

• Edited by manishcal16PPS 20 minutes ago Update