StratifiedShuffleSplit still buggy #6471
Description
y = [0, 1, 2, 3] * 3 + [4, 5] * 5
X = np.ones_like(y)
splits = StratifiedShuffleSplit(n_iter=1, train_size=11, test_size=11, random_state=0)
train, test = next(iter(splits.split(X=X, y=y)))
assert_array_equal(np.intersect1d(train, test), [])
assert_equal(len(train), 11)
assert_equal(len(test), 11)raise self.failureException('12 != 11')
this is a follow-up on #6379 and a sign that the logic is too complex. I'm doing a rewrite to fix this bug but it is not as clean as I'd like.