dask dataframe alias isna to isnull.

[TomAugspurger]

Pandas prefers isna now (not sure if isnull is being deprecated).

>>> pd.Series([1, 2, 3]).isna()

May be good to have a top-level dd.isna method as well.

[cr458]

Could have a look at implementing this? I guess it would look something like:

def isna(values):
    import pandas as pd
    return elemwise(pd.isna, values, dtype='bool')

in routines.py?

[TomAugspurger]

I think to_datetime will be a good guide for the top-level pd.isna

dask/dask/dataframe/core.py

Lines 4113 to 4116 in 4ad9622

	@wraps(pd.to_datetime)
	def to_datetime(arg, **kwargs):
	meta = pd.Series([pd.Timestamp('2000')])
	return map_partitions(pd.to_datetime, arg, meta=meta, **kwargs)

And isnull / notnull should be good guides for the methods on Series / DataFrame

dask/dask/dataframe/core.py

Lines 1499 to 1501 in 4ad9622

	@derived_from(pd.DataFrame)
	def isnull(self):
	return self.map_partitions(M.isnull)