fillna()函数

源码：

1     def fillna(self, value=None, method=None, axis=None, inplace=False,
2                limit=None, downcast=None, **kwargs):
3         return super(DataFrame,
4                      self).fillna(value=value, method=method, axis=axis,
5                                   inplace=inplace, limit=limit,
6                                   downcast=downcast, **kwargs)
7 
8     @Appender(_shared_docs['shift'] % _shared_doc_kwargs)

method : {‘backfill’, ‘bfill’, ‘pad’, ‘ffill’, None}

method为ffill时，表示dataframe中每一列向下填充，即

1 df = pd.DataFrame(  [[np.nan,2,np.nan,0],
2                      [3,4,88,1],
3                      [np.nan,np.nan,np.nan,5],
4                      [np.nan,3,np.nan,4]],
5                      columns=list('ABCD'))
6 print(df)
7 print(df.fillna(method='ffill'))

输出：

 1      A    B     C  D
 2 0  NaN  2.0   NaN  0
 3 1  3.0  4.0  88.0  1
 4 2  NaN  NaN   NaN  5
 5 3  NaN  3.0   NaN  4
 6      A    B     C  D
 7 0  NaN  2.0   NaN  0
 8 1  3.0  4.0  88.0  1
 9 2  3.0  4.0  88.0  5
10 3  3.0  3.0  88.0  4

参考：https://www.cnblogs.com/sunbigdata/p/7895295.html