def testDataFrameCorr(self):
rs = np.random.RandomState(0)
raw = rs.rand(20, 10)
raw = pd.DataFrame(np.where(raw > 0.4, raw, np.nan),
columns=list('ABCDEFGHIJ'))
raw['k'] = pd.Series(['aaa'] * 20)
df = DataFrame(raw)
result = df.corr()
pd.testing.assert_frame_equal(
self.executor.execute_dataframe(result, concat=True)[0],
raw.corr())
result = df.corr(method='kendall')
pd.testing.assert_frame_equal(
self.executor.execute_dataframe(result, concat=True)[0],
raw.corr(method='kendall'))
df = DataFrame(raw, chunk_size=6)
with self.assertRaises(Exception):
self.executor.execute_dataframe(df.corr(method='kendall'),
concat=True)
result = df.corr()
pd.testing.assert_frame_equal(
self.executor.execute_dataframe(result, concat=True)[0],
raw.corr())
result = df.corr(min_periods=7)
pd.testing.assert_frame_equal(
self.executor.execute_dataframe(result, concat=True)[0],
raw.corr(min_periods=7))
def test_dataframe_corr(setup):
rs = np.random.RandomState(0)
raw = rs.rand(20, 10)
raw = pd.DataFrame(np.where(raw > 0.4, raw, np.nan),
columns=list('ABCDEFGHIJ'))
raw['k'] = pd.Series(['aaa'] * 20)
df = DataFrame(raw)
result = df.corr()
pd.testing.assert_frame_equal(result.execute().fetch(), raw.corr())
result = df.corr(method='kendall')
pd.testing.assert_frame_equal(result.execute().fetch(),
raw.corr(method='kendall'))
df = DataFrame(raw, chunk_size=6)
with pytest.raises(Exception):
df.corr(method='kendall').execute()
result = df.corr()
pd.testing.assert_frame_equal(result.execute().fetch(), raw.corr())
result = df.corr(min_periods=7)
pd.testing.assert_frame_equal(result.execute().fetch(),
raw.corr(min_periods=7))