Series functions with level parameter produces Series with sorted index while Pandas doesn't do this

[amyskov]

System information

• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Linux Ubuntu 18.04
• Modin version (modin.__version__): 0.7.3+197.g8602fc0
• Python version: 3.8.3
• Code we can use to reproduce:

import pandas
import numpy as np

from modin.pandas.test.utils import test_data
from modin.pandas.test.test_series import create_test_series

s_modin, s_pandas = create_test_series(test_data["dense_nan_data"])
rows_number = len(s_modin.index)
level_0 = np.random.choice([x for x in range(10)], rows_number)
level_1 = np.random.choice([x for x in range(10)], rows_number)
index = pandas.MultiIndex.from_arrays([level_0, level_1])

s_modin.index = index
s_pandas.index = index

print("s_modin.all(level=0).index", s_modin.all(level=0).index) # Int64Index([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], dtype='int64')
print("s_pandas.all(level=0).index", s_pandas.all(level=0).index) # Int64Index([1, 5, 8, 2, 4, 7, 9, 3, 6, 0], dtype='int64')

print("s_modin.any(level=0).index", s_modin.any(level=0).index) #  Int64Index([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], dtype='int64')
print("s_pandas.any(level=0).index", s_pandas.any(level=0).index) # Int64Index([1, 5, 8, 2, 4, 7, 9, 3, 6, 0], dtype='int64')

print("s_modin.mad(level=0).index", s_modin.mad(level=0).index) # Int64Index([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], dtype='int64')
print("s_pandas.mad(level=0).index", s_pandas.mad(level=0).index) # Int64Index([1, 5, 8, 2, 4, 7, 9, 3, 6, 0], dtype='int64')

Describe the problem

Source code / logs