Weird behavior of IBP_RAGGR (rolling aggregation )

[sreeram18]

Hello,

I am seeing a behavior in rolling aggregation using IBP_RAGGR, where some combinations are acting in a different way than expected. I am aggregating the last 9 months of value in a KF using IBP_RAGGR. Though it is working fine with most of the combinations, I am facing issues with some as shown in the image attached

screenshot-2023-12-05-105205.png

The green highlight shows where it is working correctly , since l have value only in the previous 1 month, it is correctly aggregating that in the correct bucket (eg, I have value in M10 only and that is coming as aggregated value in M11).
Summation is happening correctly for all buckets, for combinations where it is coming correctly.

However for some combinations , like on the red and yellow highlights, though the aggregated value is correct, it is showing up 1 month after the expected bucket - eg I have value in M10, but instead of aggregating that at M11, it is getting aggregated at M12.

Any idea why this could happen?

Aggregation formula used - IBP_RAGGR("KFTOBEAGGREGATED@MTHPRODGRPLOCCUST",''SUM'',-9,9,''PASTCURRENTFUTURE'')

[Irmi_Kuntze]

Hi

When you read the Online Docu on IBP_CAGGR, you will find the note

"For the cumulative aggregation to calculate correct values, the input key figure must have values for all time periods to be aggregated.

Make sure that key figure values exist for all periods to be aggregated. If this is not the case, upload NULL values for the periods where key figure values are missing."

The Calculation stops at NULL values. Please try to have a KF inbetween such as:

KFEXTRA@MTHPRODGRPLOCCUST = JF(KFTOBEAGGREGATED@MTHPRODGRPLOCCUST, KFTOBEAGGREGATED@MTHPRODGRPLOCCUST, 0), meaning it writes a 0 where nothing would be. Then have your

IBP_RAGGR("KFEXTRA@MTHPRODGRPLOCCUST",''SUM'',-9,9,''PASTCURRENTFUTURE'')

[sreeram18]

Thanks for your suggestions, I have tried out as you mentioned - but that has not solved the issue.
(But that was an important caution in the documentation that I missed)