def test_quartile_numeric_object(self):
# Test numeric values within an object typed column.
df = pandas.DataFrame({'a': [1.0] * 9 + [2.0]}, dtype='object')
expected = df.ix[9:, :]
expected.loc[:, 'meas_flag'] = ['extreme']
expected.loc[:, 'meas_method'] = ['quartile']
result = outliers(df, 'a', 'quartile')
self.assertEqual(expected.to_string(), result.to_string())
self.assertTrue(all(['object', 'object', 'object'] == result.dtypes))
def test_quartile_none(self):
# Test a data frame containing a missing value.
df = pandas.DataFrame({'a': [1.0] * 8 + [None, 2.0]})
expected = df.ix[9:, :]
expected.loc[:, 'meas_flag'] = ['extreme']
expected.loc[:, 'meas_method'] = ['quartile']
self.assertEqual(
expected.to_string(),
outliers(df, 'a', 'quartile').to_string())
def test_quartile_empty(self):
# Test calling quartile on an empty DataFrame.
df = ec.query_measurements_original({'company': 'Company-1',
'site': 'Site-1',
'device_group': '1000',
'tester': 'Station-NONE'}, # No data for Station-NONE
ts[0],
ts[14] + timedelta(seconds=0.5)).toPandas()
self.assertEqual(0,
outliers(df, 'meas_value', 'quartile').shape[0])
def test_quartile(self):
df = ec.query_measurements_original(
{
'company': 'Company-1',
'site': 'Site-1',
'device_group': '1000',
'tester': 'Station-2'},
ts[0],
ts[14] +
timedelta(
seconds=0.5)).toPandas()
expected = df.ix[10:14, :]
expected.loc[
:,
'meas_flag'] = [
'mild',
'extreme',
'mild',
'extreme']
expected.loc[:, 'meas_method'] = ['quartile']
self.assertEqual(expected.to_string(),
outliers(df, 'meas_value', 'quartile').to_string())
def test_invalid_method(self):
df = pandas.DataFrame({'a': [1.0]})
with self.assertRaises(ValueError):
outliers(df, 'meas_value', 'INVALID_METHOD')
def test_quartile_invalid_column(self):
df = pandas.DataFrame({'a': [1.0]})
with self.assertRaises(KeyError):
outliers(df, 'INVALID_COLUMN_NAME', 'quartile')
def test_quartile_string(self):
df = pandas.DataFrame({'a': [1.0] * 8 + ['A STRING']})
self.assertEqual(0,
outliers(df, 'a', 'quartile').shape[0])