Showing posts with label partitions. Show all posts
Showing posts with label partitions. Show all posts

Friday, March 30, 2012

Problems with Switching partitions

Hi,

Im trying to switch a partition from partitioned table to a non partitioned table. I am getting the following error.

eCertified.dbo.Carrier_Trip_AvailabilitySwitch.PK_Carrier_Trip_AvailabilitySwitch_1' is in filegroup 'PRIMARY' and partition 24 of index 'eCertified.dbo.Carrier_Trip_Availability_Partition.PK_Carrier_Trip_Availability_Partition' is in filegroup 'SECONDARY24'. [SQLSTATE 42000] (Error 4939). The step failed.

The table one has 24 partitions and it has a partitioned clustered index. PK_Carrier_Trip_AvailabilitySwitch is the primary key in the non partitioned table and its is in the primary file group while the PK_Carrier_Trip_Availability_Partition is the primary key iin the partitioned table and its in the file group secondary 24.

Please let me know how to solve this issue.

When you switch a partition of a partitioned table to a non-partitioned table or a partition of another partitioned table, the source partition and the target table/partition must be in the same file group.

Monday, February 20, 2012

Problems when processing cube partitions

Hello all

I am experiencing problems with one of my cubes. We have a historical cube which is partitiones in quarters from 2004-2006, and then monthly for 2007. I have processed the dimensions in full, and now try to process only the newest cube partition, the one for june 2007, and after thats done the cube reults are quite bad, my measures are way way way off... What am I missing in here?

After processing a dimension in full mode all fact data in the cube is deleted! So now, you only have the data for June 2007.

Hope this clarifies you question.

|||Thanks for your reply, this answers my question... I wonder what to do then, as the process update takes almost double time as the process full Sad|||

Are you actually updating attributes of existing members or only adding new members?

If you are only adding new ones you could look into doing a processAdd.

|||Unfortunately every friday the dimension is "cleaned" which means removing some data from it