What is the intended outcome here?

My approach would be to incrementally append data into a table in Celonis until all required rows are captured. This ensures a complete dataset while preserving existing records.

Thanks for your reply ​@abhishek.chatu14. 

The List Items - Sharepoint Online module outputs a series of bundles (max. 3500). The array aggregator then merges these bundles into one single bundle and later the data is written into a table in Celonis. This is already implemented.

Instead of retrieving only the first 3500 items from the SahrePoint list, I want to retrieve all items (more than 3500, but the exact number is unknown as the data in the list changes). I was wondering if I can do this with a module like the repeater module that retrieves the first 3500 items from Sharepoint, in the next pass the next 3500 items and so on, which are then merged into a single bundle and written to a table. But I'm not sure how to implement it.

A more effective approach would be to leverage the Machine Learning Workbench and develop the API using Python. Action Flows may not be the best fit for your requirements, so you might want to explore alternative solutions. You could also consider seeking guidance from the Make community: Make Community.