Question

我有像这样的pandas数据框

>>> df['d']
date
2013-11-18    0
2013-11-19    0
2013-11-20    0
2013-11-21   -1
2013-11-22    0
2013-11-25    0
2013-11-26    0
2013-11-27    0
2013-11-29    0
2013-12-02    1
2013-12-03    0
2013-12-04    0
2013-12-05    0
2013-12-06    0
2013-12-09    0
2013-12-10    0
2013-12-11    0
2013-12-12    0
2013-12-13   -1
2013-12-16    0
2013-12-17    0
2013-12-18    0
2013-12-19    0
2013-12-20    0
2013-12-23    0
2013-12-24    0
2013-12-26    0
2013-12-27    0
2013-12-30    0
2013-12-31    1
2014-01-02    0
2014-01-03    0
2014-01-06    0
2014-01-07    0
2014-01-08   -1
Name: d, dtype: int64

我想将其更改为

>>> df['d']
date
2013-11-18    0
2013-11-19    0
2013-11-20    0
2013-11-21   -1
2013-11-22   -1
2013-11-25   -1
2013-11-26   -1
2013-11-27   -1
2013-11-29   -1
2013-12-02    1
2013-12-03    1
2013-12-04    1
2013-12-05    1
2013-12-06    1
2013-12-09    1
2013-12-10    1
2013-12-11    1
2013-12-12    1
2013-12-13   -1
2013-12-16   -1
2013-12-17   -1
2013-12-18   -1
2013-12-19   -1
2013-12-20   -1
2013-12-23   -1
2013-12-24   -1
2013-12-26   -1
2013-12-27   -1
2013-12-30   -1
2013-12-31    1
2014-01-02    1
2014-01-03    1
2014-01-06    1
2014-01-07    1
2014-01-08   -1
Name: d, dtype: int64

到目前为止，我使用此代码进行更改

>>> for i in range(len(df)):
...     if i != 0 and df['d'][i] ==0:
...             df['d'][i] = df['d'][i-1]
...

代码效率不高，有没有更有效的方法呢？我会假设必须有一些特殊的函数（例如apply，roll_apply）来迭代这些值，但我无法弄清楚这一点。任何帮助将不胜感激。

Answer 1

只需你可以做到这一点

df['d'].replace(0 , method = 'ffill')

如何更有效地修改pandas dataframe列

1 个答案: