Python: 当 groupby 将 dataframe 聚合到一行时，mean() 不起作用

Question

我有数据框：

time_to_rent = {'rentId': {0: 43.0, 1: 87.0, 2: 140.0, 3: 454.0, 4: 1458.0}, 'creditCardId': {0: 40, 1: 40, 2: 40, 3: 40, 4: 40}, 'createdAt': {0: Timestamp('2020-08-24 16:13:11.850216'), 1: Timestamp('2020-09-10 10:47:31.748628'), 2: Timestamp('2020-09-13 15:29:06.077622'), 3: Timestamp('2020-09-24 08:08:39.852348'), 4: Timestamp('2020-10-19 08:54:09.891518')}, 'updatedAt': {0: Timestamp('2020-08-24 20:26:31.805939'), 1: Timestamp('2020-09-10 20:05:18.759421'), 2: Timestamp('2020-09-13 18:38:10.044112'), 3: Timestamp('2020-09-24 08:53:22.512533'), 4: Timestamp('2020-10-19 17:10:09.110038')}, 'rent_time': {0: Timedelta('0 days 04:13:19.955723'), 1: Timedelta('0 days 09:17:47.010793'), 2: Timedelta('0 days 03:09:03.966490'), 3: Timedelta('0 days 00:44:42.660185'), 4: Timedelta('0 days 08:15:59.218520')}}

按列 'creditCardId' 聚合数据框并具有 'rent_time' 平均值的想法。理想的输出应该是：

creditCardId        rent_time mean
40                  0 days 05:08:10.562342

如果我运行代码：

print (time_to_rent['rent_time'].mean())

它工作正常，我有“0 天 05:08:10.562342” 作为输出。但是当我试图通过以下方式分组时：

time_to_rent.groupby('creditCardId', as_index=False)[['rent_time']].mean()

我收到错误信息：

~\anaconda3\lib\site-packages\pandas\core\groupby\generic.py in _cython_agg_blocks(self, how, alt, numeric_only, min_count)
   1093 
   1094         if not (agg_blocks or split_frames):
-> 1095             raise DataError("No numeric types to aggregate")
   1096 
   1097         if split_items:

DataError: No numeric types to aggregate

如果我使用命令：

time_to_rent = time_to_rent.groupby('creditCardId', as_index=False)[['rent_time']]

它 returns 只有“”

你能帮我理解我的错误在哪里吗？

Answer 1

这不是你的错误，可能是 Pandas 中的错误，因为 Timedelta 可以求平均。 A work-around 是 apply:

time_to_rent.groupby('creditCardId')['rent_time'].apply(lambda x: x.mean())

输出：

creditCardId
40   0 days 05:08:10.562342200
Name: rent_time, dtype: timedelta64[ns]

Python: 当 groupby 将 dataframe 聚合到一行时，mean() 不起作用

Python: mean() doesn't work when groupby aggregates dataframe to one line

python

grouping

mean

pandas