Possibility to read ROOT files into Dask DataFrames in Python

Hello,

I was wondering whether there is any interface, which gives you the possibility to read ROOT files directly into Dask DataFrames using Python.
I am aware of the functionality to read ROOT files and obtain a Pandas DataFrame. However this means, that you are usually constrained by the amount of RAM in your computer.
I am also aware of uproot, which seems to have had a Dask backend. But unfortunately this seems to have been removed with release 4.0.0rc1 of uproot.
Is there any recommendation on how something like this could be done?

Thank you for any tips

I think @eguiraud can help you.

Hi @janhein ,
and welcome to the ROOT forum!

Currently, ROOT does not offer direct integration with dask dataframes. A possible solution is to do perform a first selection of events and columns using RDataFrame, then export to NumPy arrays that subset of the data (see e.g. ROOT: tutorials/dataframe/df026_AsNumpyArrays.py File Reference) and use dask from there.

Cheers,
Enrico