def gen_unique_values(data_length, dtype='int', seed=None):
"""
data_length: result length of array of unique values,
dtype: dtype of generated array,
seed: seed to initialize random state
"""
if dtype in ('float', 'int'):
values = np.arange(data_length, dtype=dtype)
if dtype == 'str':
values = gen_strlist(data_length)
return gen_arr_from_input(data_length, values, repeat=False, seed=seed)
def gen_arr_of_dtype(data_length,
dtype='float',
random=True,
limits=None,
nunique=1000,
input_data=None,
seed=None):
"""
data_length: result array length,
dtype: dtype of generated array,
limits: a tuple of (min, max) limits for numeric arrays,
nunique: number of unique values in generated array,
input_data: 1D sequence of values used for generation of array data,
seed: seed to initialize random state
"""
if seed is not None:
np.random.seed(seed)
# prefer generation based on input data if it's provided
if input_data is not None:
return gen_arr_from_input(input_data, data_length, random=random)
if dtype == 'float':
return np.random.ranf(data_length)
if dtype == 'int':
default_limits = (np.iinfo(dtype).min, np.iinfo(dtype).max)
min_value, max_value = limits or default_limits
return np.random.randint(min_value, max_value, data_length)
if dtype == 'str':
default_strings = gen_strlist(nunique)
return np.random.choice(default_strings, data_length)
if dtype == 'bool':
return np.random.choice([True, False], data_length)
return None
def test_df_groupby_mean_by_str_sort_false(self):
self._test_df_groupby_method('mean',
usecase_name='by_str_mean',
input_data=[gen_strlist(n_groups_default, 3, 'abcdef')],
groupby_params={'sort': 'False'})