def test_groupby_to_series_to_frame_2(self): df = pd.DataFrame({'a': [6, 2, 2], 'b': [4, 5, 6]}) labels = ['g1', 'g1', 'g2'] benchmark = df.groupby(labels).apply(frame_to_series) result = pandas_easy.groupby_to_series_to_frame( df, frame_to_series, 1, use_apply=False, by=labels) assert_frame_equal(result, benchmark)
def user_appCategory(): user_app_category = get_userID_appID() user_app_category = user_app_category[:1000][:] app_category_series = user_app_category['appCategory'].unique() columns_name = [ 'user_category_count_' + str(i) for i in app_category_series.tolist() ] func = partial(count_user_appCategory, app_category_series) user_app_cateogry_dataframe = pandas_easy.groupby_to_series_to_frame( user_app_category, func, n_jobs=mp.cpu_count() - 1, use_apply=True, by='userID') ''' user_app_category_groupby_user = user_app_category.groupby('userID',as_index = False) user_app_cateogry_list = papply(user_app_category_groupby_user,count_user_appCategory,app_category_series) #user_app_cateogry_list = user_app_category_groupby_user.apply(lambda x:count_user_appCategory(x,app_category_series)) user_app_cateogry_dataframe = pd.DataFrame(list(user_app_cateogry_list), columns=['userID'] + columns_name) ''' #print(user_app_cateogry_dataframe) split_point = len(user_app_cateogry_dataframe) // 4 data1 = user_app_cateogry_dataframe[:split_point][:] data2 = user_app_cateogry_dataframe[split_point:split_point * 2] data3 = user_app_cateogry_dataframe[split_point * 2:split_point * 3] data4 = user_app_cateogry_dataframe[split_point * 3:][:] pickle.dump(data1, open(user_category_path1, 'wb'), protocol=4) pickle.dump(data2, open(user_category_path2, 'wb'), protocol=4) pickle.dump(data3, open(user_category_path3, 'wb'), protocol=4) pickle.dump(data4, open(user_category_path4, 'wb'), protocol=4)
def test_groupby_to_series_to_frame_3(self): df = pd.DataFrame({'a': [6, 2, 2], 'b': [4, 5, 6]}) benchmark = df.groupby('a').apply(max) result = pandas_easy.groupby_to_series_to_frame(df, max, 1, by='a') assert_frame_equal(result, benchmark)