Set partitions on existing index in Dask dataframe

Question

If I have an already indexed Dask dataframe with

>>> A.divisions
(None, None)
>>> A.npartitions
1

and I want to set the divisions, so far I'm doing

A.reset_index().set_index("index", divisions=sorted(divisions))

because A.repartition(divisions=sorted(divisions)) complains "left side of old and new divisions are different". Is there a better way?

Answer 1

As of dask.__version__ == '0.16.0' if you happen to know the divisions of an existing dataframe you can assign them directly.

A.divisions = tuple(divisions)