Specifically on the point about Pandas, last time I checked the inplace flag did not save any memory. Under the hood, it creates a temporary copy which is copied back to the original dataframe, and then the temporary gets removed next garbage collection cycle. If that is no longer the case I would love to know about it though!