May 15, 2003 at 5:45 am
Hi all,
I'm having the following problem...
I've just done a incremental update of my cube. After the processing I have double values i.e. the value is 10 where it should be 5. This is for all values in the cube the same.
Could there be a problem in the selection of the processing mode ?
Before the processing I deleted one shared dimension which previously had been deleted in the cube design, processed all demensions and then did an incremental update.
Does anybody seen this before ??
Thanks in advance
Regards,
Bart
May 15, 2003 at 8:14 am
You more than likely have a non-unique key value in either your fact table or dimension table. Instead of joining unique value with another unique value you have an instance where the unique value is joining with two rows in the other table.
You might also check to make sure the fact table or one of the dimension tables wasn't loaded twice by accident.
Skip Gallagher
Quest Industries, Inc.
(479) 685-4546
Skip Gallagher
Quest Industries, Inc.
(479) 366-0789
May 15, 2003 at 11:49 pm
Thanks Skip,
I think I was a bit to fast with my question.
The books online document gave me also a reason why I got the same data twice.
My fact table is based on a view and I did no joining.
Incremental Updates:
Incremental updates allow you to keep the contents of a cube current without requiring you to reprocess the cube in full when you add new data. An incremental update involves creating a temporary partition, filling it with updated source data, processing the temporary partition, and then merging it into another partition in the cube.
Data to be added to a cube can come from the original fact table or from a separate fact table with a structure identical to the original. If you add data from the original fact table, take care not to duplicate data that already exists in the cube. Set the SourceTableFilter property before processing the created temporary partition to restrict the data that is imported from the fact table. Temporary partitions created for this purpose are indicated by a tilde (~) character preceding the name of the temporary partition.
If data to be added to a cube comes from the fact table from which the cube was originally created, a risk of duplicate aggregation occurs. The cube uses the SourceTableFilter property to screen incoming data from a fact table; when performing an incremental update, it adds the aggregations computed from the fact table to the aggregations stored by the cube. If the same table is run twice, once to construct the original aggregations, and again as part of an incremental update, you will receive the same data twice, added together in the cube. The SourceTableFilter property can be used to screen out existing, already aggregated data in the fact table, preventing duplicate aggregation.
If you add data from a fact table that includes new members of a dimension, you must also reprocess the affected dimension using the processRefreshData for the Process method of the dimension object.
Maybe the above could be helpfull to you..
Skip thanks for the response.
Greetz,
Bart
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply