DataFrame level sum and mean fail due to lack of axis and skipna keyword arguments in cudf.DataFrame.sum. Raising this issue here for tracking purposes and will raise the upstream cuDF issue to resolve the issue.
import cudf
import dask_cudf
df = cudf.DataFrame({'a':range(10), 'b':range(10, 20)})
print(df.mean())
ddf = dask_cudf.from_cudf(df, 2)
ddf.sum()
ddf.mean() # fails also for the same reason
a 4.5
b 14.5
dtype: float64
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
<ipython-input-1-ccf591ee1075> in <module>
5 print(df.mean())
6 ddf = dask_cudf.from_cudf(df, 2)
----> 7 ddf.sum()
8 ddf.mean()
/conda/envs/rapids/lib/python3.7/site-packages/dask/dataframe/core.py in sum(self, axis, skipna, split_every, dtype, out, min_count)
1333 out=None, min_count=None):
1334 result = self._reduction_agg('sum', axis=axis, skipna=skipna,
-> 1335 split_every=split_every, out=out)
1336 if min_count:
1337 return result.where(self.notnull().sum(axis=axis) >= min_count,
/conda/envs/rapids/lib/python3.7/site-packages/dask/dataframe/core.py in _reduction_agg(self, name, axis, skipna, split_every, out)
1297 axis = self._validate_axis(axis)
1298
-> 1299 meta = getattr(self._meta_nonempty, name)(axis=axis, skipna=skipna)
1300 token = self._token_prefix + name
1301
TypeError: sum() got an unexpected keyword argument 'axis'
beckernick
commented
5 years ago
DataFrame level
sumandmeanfail due to lack of axis and skipna keyword arguments in cudf.DataFrame.sum. Raising this issue here for tracking purposes and will raise the upstream cuDF issue to resolve the issue.