如何获取pandas的日差

Question

例子=

day | hours 
monday  5
monday   6
tuesday  5
tuesday  6
tuesday  7
wednesday 5
wednesday 6
wednesday 7

预期结果：

day | hours
monday 1
tuesday 2
wednesday 2

Answer 1

您只想从每个组中减去 first 和 last 行。

使用Groupby.agg, pick first and last rows per group. Then, use df.diff with MultiIndex.droplevel:

In [3831]: x = (
      ...:     df.groupby("day")
      ...:     .agg({"hours": ["first", "last"]})
      ...:     .diff(axis=1)[("hours", "last")]
      ...:     .reset_index()
      ...:     .droplevel(level=1, axis=1)
      ...: )

In [3829]: x
Out[3829]: 
         day  hours
0     monday    1.0
1    tuesday    2.0
2  wednesday    2.0

Answer 2

我读得太快了，我们试试：

df.groupby('day', as_index=False)['hours'].apply(lambda x: x.max()-x.min())

或者

df.groupby('day', as_index=False)['hours'].apply(np.ptp)

输出：

         day  hours
0     monday      1
1    tuesday      2
2  wednesday      2

如何获取pandas的日差

How to obtain the daily difference in pandas

python

pandas

data-science