Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Optional 'groups' parameter in GroupKFold().split()

See original GitHub issue

In the documentation of GroupKFold().split() the parameter groups is indicated as optional.

However, the following example fails with: The 'groups' parameter should not be None.

>>> import numpy as np
>>> from sklearn.model_selection import GroupKFold
>>> X = np.array([[1, 2], [3, 4], [5, 6], [7, 8]])
>>> y = np.array([1, 2, 3, 4])
>>> group_kfold = GroupKFold(n_splits=2)
>>> for train_index, test_index in group_kfold.split(X, y):
...     print("TRAIN:", train_index, "TEST:", test_index)
... 
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/lib/python3.7/site-packages/sklearn/model_selection/_split.py", line 331, in split
    for train, test in super(_BaseKFold, self).split(X, y, groups):
  File "/usr/lib/python3.7/site-packages/sklearn/model_selection/_split.py", line 100, in split
    for test_index in self._iter_test_masks(X, y, groups):
  File "/usr/lib/python3.7/site-packages/sklearn/model_selection/_split.py", line 112, in _iter_test_masks
    for test_index in self._iter_test_indices(X, y, groups):
  File "/usr/lib/python3.7/site-packages/sklearn/model_selection/_split.py", line 504, in _iter_test_indices
    raise ValueError("The 'groups' parameter should not be None.")
ValueError: The 'groups' parameter should not be None.

Is this a documentation issue or a bug?