def __init__(
self, obj, by=None, level=None, sort=True, as_index=True, dropna=True
):
"""
Group a DataFrame or Series by a set of columns.
Parameters
----------
by : optional
Specifies the grouping columns. Can be any of the following:
- A Python function called on each value of the object's index
- A dict or Series that maps index labels to group names
- A cudf.Index object
- A str indicating a column name
- An array of the same length as the object
- A Grouper object
- A list of the above
level : int, level_name or list, optional
For objects with a MultiIndex, `level` can be used to specify
grouping by one or more levels of the MultiIndex.
sort : True, optional
If True (default), sort results by group9s). Note that
unlike Pandas, this also sorts values within each group.
as_index : bool, optional
If as_index=True (default), the group names appear
as the keys of the resulting DataFrame.
If as_index=False, the groups are returned as ordinary
columns of the resulting DataFrame, *if they are named columns*.
dropna : bool, optional
If True (default), do not include the "null" group.
"""
self.obj = obj
self._as_index = as_index
self._sort = sort
self._dropna = dropna
if isinstance(by, _Grouping):
self.grouping = by
else:
self.grouping = _Grouping(obj, by, level)
self._groupby = libgroupby.GroupBy(self.grouping.keys, dropna=dropna)
def _groupby(self):
return libgroupby.GroupBy(self.grouping.keys, dropna=self._dropna)