python如何读取数据框中的数据_python-Pandas数据框通过匹配另一个数据框中...

更新：从the comment回答以下问题：

What if A is a multiindex? With [‘bucket1′,’bucket2’] as index but we

only care for bucket1?

In [140]: A

Out[140]:

value

bucket1 bucket2

1 10 0.001855

11 0.000120

2 12 0.000042

13 0.001888

In [141]: B

Out[141]:

bucket num

0 1 0.5

1 2 0.3

In [142]: A['new'] = A.value / A.reset_index().iloc[:, 0].map(B.set_index('bucket').num).values

In [143]: A

Out[143]:

value new

bucket1 bucket2

1 10 0.001855 0.003710

11 0.000120 0.000240

2 12 0.000042 0.000140

13 0.001888 0.006293

旧答案：

In [61]: A['new'] = A.value.div(A.bucket.map(B.set_index('bucket').num))

In [62]: A

Out[62]:

bucket value new

0 1 0.001855 0.003710

1 1 0.000120 0.000240

2 2 0.000042 0.000140

3 2 0.001888 0.006293

或作为虚拟列：

In [60]: A.assign(new=A.value/A.bucket.map(B.set_index('bucket').num))

Out[60]:

bucket value new

0 1 0.001855 0.003710

1 1 0.000120 0.000240

2 2 0.000042 0.000140

3 2 0.001888 0.006293

说明：

In [65]: B.set_index('bucket')

Out[65]:

num

bucket

1 0.5

2 0.3

In [66]: A.bucket.map(B.set_index('bucket').num)

Out[66]:

0 0.5

1 0.5

2 0.3

3 0.3

Name: bucket, dtype: float64