Hi everyone,
we are currently restructuring our data architecture in Celonis. For each process area (P2P, AP, AR), we have separate data pools where we extract, transform, and load the respective data models.
In the AP area, we already use a dedicated Central Extraction Pool, which only handles the extraction of raw tables. The actual AP data pool then receives these extracted tables and performs the transformations.
Now, we want to extend this concept gradually to other areas. For example, we plan to stop extracting the CDPOS table in the Purchasing Pool and instead consume it from the Central Pool. This step-by-step migration should help us reduce redundancy and improve maintainability.
However, I have a specific question regarding this setup: If I start consuming tables from the Central Pool in the Purchasing Pool (e.g., CDPOS), what happens to other tables like EKKO, which are still being extracted directly in the Purchasing Pool?
Will my transformations in the Purchasing Pool use the locally extracted EKKO table, or will they automatically switch to the EKKO table from the Central Pool, assuming it exists there too?
I want to avoid any unintended behavior or data mismatches, especially during the transition phase where some tables are still extracted locally and others are sourced from the Central Pool.
Has anyone gone through a similar migration? How does Celonis determine which version of a table is used when the same table exists in both the local and central pools?
Thanks in advance for your insights!
