I am using mstrio to push a huge dataframe to MSTR. Every day 1/4 (the newest x days) of the dataset gets an update (row's are deleted, updated or added). There are different update strategies to push the data to mstr, being 'update', 'upsert', or 'replace'. The problem with 'update' and 'upsert' is, that it cannot really handle deleted rows, which is key in our application. 'Replace' is doing the job, but the problem is that it deletes all data, so every day we have to push the complete dataset to mstr again.
So optimally I would only use the update strategy 'replace' for my latest x days, and use upsert strategy for the older data. Is there a way to do this?