cancel
Showing results for 
Search instead for 
Did you mean: 

Overwrite feature in Cube

Former Member
0 Kudos

Hi Gurus,

I have an issue with duplicate records, the scenario is as follows,

I have a cube which stores historical data for sales calls made.It is fed by an ODS everyday via a delta load. Now, this cube has a characteristics Category.

The Cube has a record with category "A" , the user changes this category from "A" to "X" in R/3.When the delta is loaded to the cube , this change is added as another record and when we execute the report , the user sees two records for the same calls made. In Actual the user should only see one record with this new category"X".

Is there any way to overwrite the old category with the new category during delta load into the cube ? so that I can see only one record instead of two records.

Thanks in advance...

Accepted Solutions (0)

Answers (2)

Answers (2)

Former Member
0 Kudos

Hi Narendra,

You can try to duplicate record with 'X' in Start Routine of Update Roules and change 'X' with 'A' for the second record and introduce 'D' for field 0RECORDMODE that you can introduce in your Cube. In this way you can delete in Cube the record with 'A'.

Logically all other fields must be equal to the old record uploaded with 'A' in the past.

Ciao.

Riccardo.

Former Member
0 Kudos

Probably that is because category is a key fields in ODS; if you change the ods without category in key fieds, the delta should generate two records, one with cat. A and all KF wit sign inversion, and one with CAT. X.

Hope it helps.

Regards

Former Member
0 Kudos

Hi Paolo,

Thanks for the response , if two records are generated then it should have two records in the CUBE right?

I need only one record with category "X" which should overwrite "A".

Thanks in advance..

Former Member
0 Kudos

Hi Narendra,

I am going to explain Paolo suggestion:

If you put Category IO not in Key Fields of ODS but in DataFields when uploading in ODS the second record, with 'X' value, will overwrite first record, with 'A' value. Due to this overwrite in ODS the system will produce two records for delta upload from ODS to Cube: one record with 'X' value and with all Key Figures with correct value and sign, one record with 'A' value and with all Key Figures with correct value and inverted sign. When these two records are uploaded into cube you will never see 'A' record in BEx, because all Key Figures values are equal '0' now.

Logically this only an explication of Paolo suggestion, so for points I consider correct to assign these to Paolo if you are satisfied.

Ciao.

Riccardo.

Former Member
0 Kudos

Hi Riccardo,

Thanks for the explanation, but i have the IO in the data field and its a direct map into the cube.

My apologies for mentioning that Category is a Data field in the ODS.

Thanks

Former Member
0 Kudos

Hi Narendra,

If Category is already a DataField and you find two records in Cube you must have in Key fields another IO of Type char that determine the split into two records: you have two different values not only for A and X but there is another field that different values beetween first and second upload, maybe a date.

In this case the problem is a bit more complex. I think that the only solution is to exclude from query the IO is producing the duplicate records and during upload make the duplication/inversion of sign/change of X to A I was suggesting you in previuos message, also without the D value in RecordMode.

If you need more detailed information and an example please write me to riccardoventu@virgilio.it introducing also a full record with A and its subsequent record with X.

Ciao.

Riccardo.