ODS Issue

[Former Member]

Hi All,

<b>CASE</b>: We are loading the datafile from non sap system into BW for a file called Billing datafile(ascii).

Our objective is to build Billing Cube from Billing ODS.

<b>situation:</b> Billing data file is loaded in the Billing ODS.

Billing Update data file is then loaded, where each row leads to update of billing ODS row already loaded.

Billing Cube is then built from data in Billing ODS

<b>Problem facing </b>:

Bill cube shows duplicate rows for a billing line item (where single row is expected) with only one field’s value repeating and all other field values being 0 in duplicate record.ODS shows single row for the billing item in question.If Billing cube is dropped and re-created, duplicate rows go away.

What is the possible best solution to avoid duplicate records and to load all other fields values.

Points will be assigned on reply

[Former Member · Accepted solution]

Hi,

How about building an aggregate?

Cheers,

Gimmo

[Former Member · Accepted solution]

Hi,

Assume that your cube has :

Documentnumber and Netsales.

And Asume that you had a record in the Cube and ODS like

100000 100

And assume That the above record is changed in ODS as like below:

100000 200

So the delta records available for the delta uload to Cube from ODS are:

100000 -100

100000 200

So after this delta load to Cube you will have the records in the cube as follws:

100000 0

100000 200

With rgds,

Anil Kumar Sharma .P

[Former Member]

Mahesh,

In that case, there will be two records for 100 and 100 instead of 0 and 200.

Request id 1 100000 100

Request id 2 100000 +100 (summation of -100 and +200).

What am I missing?

Abdul

[Former Member]

Question or rather a clarification from Anil Kumar Sharma on one of the earlier posts in this thread.

Anil,

In your example, you have stated the following:

"So after this delta load to Cube you will have the records in the cube as follws:

100000 0

100000 200"

Assuming that the document number is the only characteristic in the cube, would there be one row or two rows after the negative and the positive amounts are applied?

Because if the document number is the key for the fact table, how can the fact table contain two duplicate rows for the same key? Am I safe to assume there will be only one row for "100000 200" instead of the two for 0 and 200. Just curious.

Thanks.

Abdul

[Former Member]

Abdul,

If I have not mistaken, Anil's explanation is based on the Package dimension.

I think Anil meant that the first record is coming with the Request ID from the first load while the second record has it's own Request ID.

Unless these two requests are compressed, fact table will have 2 different rows.

[Former Member]

Hi Kim,

It seems that you are loading "full update" mode from ODS to infocube. If this is the case then you need to delete the entire data from infocube and then reload the data from ODS to infocube with "initial update" and in the next load select to "Delta update". Using this your problem will be solved.

Hope this helps you!

SAM

[former_member184494]

Kim,

Welcome to SDN,

1. When loadin to a cube , all fields ( characteristics in the update rules ) are key fields.

2. Anil kumar sharma has given a very good explanation of how the records get stored in the cube

3. Please view the contents of the cube either by request ( if you do not want to see any report ) or through a report , the records would automatically get summarized (0+200 is still 200)

4. Also if you do not want to see the duplicate rows , compress the cube.

My 0.02

Arun

Assign points if useful

[Former Member]

Is the load to the cube a full load or a delta?