python数据帧_将列从一个数据帧添加到python中的数据帧

I have two data frames in python. The first is raw rainfall data for a single day of year and the second is the sum of daily rainfall using group.by.

One data frame looks like this (with many more rows in between device_ids):

>>> df1

device_id rain day month year

0 9z849362-b05d-4317-96f5-f267c1adf8d6 0.0 31 12 2016

1 9z849362-b05d-4317-96f5-f267c1adf8d6 0.0 31 12 2016

6 e7z581f0-2693-42ad-9896-0048550ccda7 0.0 31 12 2016

11 e7z581f0-2693-42ad-9896-0048550ccda7 0.0 31 12 2016

12 ceez972b-135f-45b3-be4w-7c23102676bq 0.2 31 12 2016

13 ceez972b-135f-45b3-be4w-7c23102676bq 0.0 31 12 2016

18 ceez972b-135f-45b3-be4w-7c23102676bq 0.0 31 12 2016

19 1d28dz3a-c923-4967-a7bb-5881d232c9a7 0.0 31 12 2016

24 1d28dz3a-c923-4967-a7bb-5881d232c9a7 0.0 31 12 2016

25 a044ag4f-fd7c-4ae4-bff3-9158cebad3b1 0.0 31 12 2016

29 a044ag4f-fd7c-4ae4-bff3-9158cebad3b1 0.0 31 12 2016

29 a044ag4f-fd7c-4ae4-bff3-9158cebad3b1 0.0 31 12 2016

... ... ... ... ... ...

3903 9z849362-b05d-4317-96f5-f267c1adf8d6 0.0 31 12 2016

3904 9z849362-b05d-4317-96f5-f267c1adf8d6 0.0 31 12 2016

3905 9z849362-b05d-4317-96f5-f267c1adf8d6 0.0 31 12 2016

And the other looks something like this:

>>> df2

rain

device_id

1d28dz3a-c923-4967-a7bb-5881d232c9a7 0.0

9z849362-b05d-4317-96f5-f267c1adf8d6 0.0

a044ag4f-fd7c-4ae4-bff3-9158cebad3b1 1.2

ceez972b-135f-45b3-be4w-7c23102676bq 2.2

e7z581f0-2693-42ad-9896-0048550ccda7 0.2

... which I got by using:

df2 = df1.groupby(['device_id'])[["rain"]].sum()

I want my final data frame to look like this:

>>> df3

rain day month year

device_id

1d28dz3a-c923-4967-a7bb-5881d232c9a7 0.0 31 12 2016

9z849362-b05d-4317-96f5-f267c1adf8d6 0.0 31 12 2016

a044ag4f-fd7c-4ae4-bff3-9158cebad3b1 1.2 31 12 2016

ceez972b-135f-45b3-be4w-7c23102676bq 2.2 31 12 2016

e7z581f0-2693-42ad-9896-0048550ccda7 0.2 31 12 2016

Which is to say that I want the "day month year" columns from df1 to be added to df2. I'm not sure if I should use merge, append, or do something else.

解决方案

Maybe this will work? groupby day month and year as well.

df.groupby(['device_id', 'day', 'month', 'year']).sum()

rain

device_id day month year

1d28dz3a-c923-4967-a7bb-5881d232c9a7 31 12 2016 0.0

9z849362-b05d-4317-96f5-f267c1adf8d6 31 12 2016 0.0

a044ag4f-fd7c-4ae4-bff3-9158cebad3b1 31 12 2016 0.0

ceez972b-135f-45b3-be4w-7c23102676bq 31 12 2016 0.2

e7z581f0-2693-42ad-9896-0048550ccda7 31 12 2016 0.0

Or you could add reset_index to return these columns to the DataFrame like

df.groupby(['device_id', 'day', 'month', 'year']).sum().reset_index()

0 1d28dz3a-c923-4967-a7bb-5881d232c9a7 31 12 2016 0.0

1 9z849362-b05d-4317-96f5-f267c1adf8d6 31 12 2016 0.0

2 a044ag4f-fd7c-4ae4-bff3-9158cebad3b1 31 12 2016 0.0

3 ceez972b-135f-45b3-be4w-7c23102676bq 31 12 2016 0.2

4 e7z581f0-2693-42ad-9896-0048550ccda7 31 12 2016 0.0

Or the following should match your index / column structure exactly.

df.groupby(['device_id', 'day', 'month', 'year']).sum().reset_index([1, 2, 3])