Python - Groupby DataFrameGroupBy 对象

Question

我有一个 Python 中的 panda 数据框，我正在其中应用 groupby。然后我想对之前的结果应用新的 groupby + sum 。更具体地说，首先我正在做:

  check_df = data_df.groupby(['hotel_code', 'dp_id', 'market', 'number_of_rooms'])
[['market', 'number_of_rooms']]

然后我想做:

check_df = check_df.groupby(['market'])['number_of_rooms'].sum()

因此，我收到以下错误:

    AttributeError: Cannot access callable attribute 'groupby' of 'DataFrameGroupBy'
objects, try using the 'apply' method

我的初始数据如下所示:

hotel_code | market | number_of_rooms | ....
---------------------------------------------
001        |   a    |      200        | ... 
001        |   a    |      200        |
002        |   a    |      300        | ...

请注意，我可能有像 (a - 200) 这样的对的重复项，这就是为什么我需要第一个 groupby。我最终想要的是这样的:

Market | Rooms
--------------
a      | 3000
b      | 250

我只是想将以下 sql 查询转换为 python:

select a.market, sum(a.number_of_rooms)
from (
        select market, number_of_rooms
        from opinmind_dev..cg_mm_booking_dataset_full
        group by hotel_code, market, number_of_rooms
        ) as a
group by market ;

有什么想法可以解决这个问题吗？如果您需要更多信息，请告诉我。

ps。我是 Python 和数据科学新手

Answer 1

IIUC，而不是:

check_df = data_df.groupby(['hotel_code', 'dp_id', 'market', 'number_of_rooms'])
[['market', 'number_of_rooms']]

你应该简单地这样做:

check_df = data_df.drop_duplicates(subset=['hotel_code', 'dp_id', 'market', 'number_of_rooms'])\
                  .loc[:, ['market', 'number_of_rooms']]\
                  .groupby('market')\
                  .sum()