Hi all, QQ on memory datasets in pipeline. I though in previous versions of Kedro <0.19.8, Kedro automatically created assumed a MemoryDatset when having the same input as the previous node's output? However, now it seems I need to explicitly declare in the DataCatalog. Is there a way to get the old behavior back? Cheers, Thomas

Hi Thomas, that was certainly the case and it still is as far as I can tell from running a quick test. Can you post some more info about why it seems that your data isn't saved in

MemoryDataset

anymore?

I don’t think this was changed - implicit creation of MemoryDatasets without the need to declare them in catalog is one of the key things in Kedro. If that stops working, all users would immediately report the issue I think… So I’d assume it works as before, and would investigate into how your pipelines are registered and how you run them.