Closed
Description
It would be nice to support sparse arrays with dtype
that is not just float64
. Specifically, I have some data that is int8
but sparse, it would be nice to be able to create a SparseArray
with the SparseIndex
feature tracking which values are set and which are not set.
Obviously, np.nan
can't be used to indicate not a value, but maybe a default can be provided (in my case, zero is a nice default).