Pull function –> how many PO items are created for one PR

Hey Oli, Disregarding the error you are getting, the way the EBAN Table is structured (Primary Key: MANDT, BANFN, BNFPO) and the way it references to the EKPO Table (MANDT, EBELN, EBELP --> Primary Key EKPO) it is not possible for one EBAN item to point to more than one entry in the EKPO, thus meaning it is a 1:1 OR N:1 (EBAN:EKPO) relationship. So there could possibly be 1 or more PRs for one PO, but not 1 or more POs for one PR. The is only one, very unlikely, case that in your system the EBAN reference only to EKKO by leaving the EBELP entry empty, then you just need to count the entries for the respective EBELN. What you maybe mean and looking for is how an EKPO entry can interact with the combination of AUFK / AUPO and that a single EKPO entry can be split to multiple entries, or the other way around, how many small orders can be consolidated to one bigger order. I hope I could clarify the topic a bit, Best Regards, Benedict

Thanks for your reply Benedict! You are right, there shall be no cases where one PR item is referenced in different POs. However, those cases occur unfortunately anyway from time to time in our SAP systems and thats the reason why we want to analyze those undesired cases with Celonis. The SQL statement to analyze those cases would probably look something like that: Select * for EBAN - EBELP Where Count of (EKPO - BANFN = EBAN - EBELN & EKPO - BNFPO = EBAN - EBELP) > 1 Can you help me translating that to PQL so that I can analyze based on PR item level the cases which have a 1:N relationship (EBAN:EKPO). Thanks in advance, Oli

Hey Oli, The EKPO.BANFN never (logically correctly) equals the EBAN.EBELN field, its either EKPO.BANFN = EBAN.BANFN or EKPO.EBELN = EBAN.EBELN. Id recommend Creating an OLAP Table with these informations: Dimensions: EBAN.BANFN EBAN.BNFPO KPI: COUNT_TABLE(EKPO) And then sort by the KPI descending and see if there are actually cases, that match your thought process. What also would work, is using only EKPO, like this: Dimensions: EKPO.BANFN EKPO.BNFPO KPI: COUNT_TABLE(EKPO) Its basically like a SQL script saying SELECT BANFN, BNFPO, COUNT(1) FROM EKPO GROUP BY BANFN, BNFPO Best Regards, Benedict

Thanks - the second option (using only EKPO fields) actually works and shows exactly the cases we were looking for. Thanks for the always fast and great support