Numpy 中的通用函数方法（逐元素数组方法）对 pandas 也有效

In [54]: frame = pd.DataFrame(np.random.randn(4, 3), columns=list('bde'),
    ...:                      index=['Utah', 'Ohio', 'Texas', 'Oregon'])

In [55]: frame
Out[55]:
               b         d         e
Utah    0.053371 -2.529817  0.368989
Ohio   -1.696226 -0.244630 -0.470242
Texas   2.574698 -2.013593 -0.644810
Oregon -0.739534  0.488870  1.704059

In [56]: np.abs(frame)
Out[56]:
               b         d         e
Utah    0.053371  2.529817  0.368989
Ohio    1.696226  0.244630  0.470242
Texas   2.574698  2.013593  0.644810
Oregon  0.739534  0.488870  1.704059

将一个函数用到一行或一列一维数组上

In [57]: f = lambda x: x.max() - x.min()

In [58]: frame.apply(f)
Out[58]:
b    4.270924
d    3.018687
e    2.348869
dtype: float64

每一列调用一次

In [59]: frame.apply(f, axis='columns')
Out[59]:
Utah      2.898806
Ohio      1.451596
Texas     4.588292
Oregon    2.443593
dtype: float64

每一行调用一次

In [60]: def f(x):
    ...:     return pd.Series([x.min(), x.max()], index=['min', 'max'])
    ...:

In [61]: frame.apply(f)
Out[61]:
            b         d         e
min -1.696226 -2.529817 -0.644810
max  2.574698  0.488870  1.704059

逐元素的 Python 函数也可使用，如 applymap 方法

In [62]: format = lambda x: '%.2f' % x

In [63]: frame.applymap(format)
Out[63]:
            b      d      e
Utah     0.05  -2.53   0.37
Ohio    -1.70  -0.24  -0.47
Texas    2.57  -2.01  -0.64
Oregon  -0.74   0.49   1.70

之所以命名为 applymap ，是因为 Series 有 map 方法

In [64]: frame['e'].map(format)
Out[64]:
Utah       0.37
Ohio      -0.47
Texas     -0.64
Oregon     1.70
Name: e, dtype: object