concat

Pandas提供了把多个DataFrame合并链接成一个DataFrame的concat的方法：

In [2]: import pandas as pd

In [3]: import numpy as np

In [4]: data = pd.DataFrame(np.random.randn(10, 4))

In [5]: data
Out[5]:
          0         1         2         3
0 -0.150377  0.473775 -0.564428  1.867808
1  0.178880 -0.356651 -0.864143 -0.325870
2  0.635222 -0.502338 -0.702672  1.469388
3 -0.921042 -0.143417  0.629042 -1.312538
4  0.224232  0.414366 -0.575869 -1.002100
5 -1.155228 -0.739624  0.131099 -1.037161
6 -1.782056 -0.316029 -0.005173  1.159687
7  0.878195  0.436940 -0.048127 -1.952570
8  1.511242 -0.189323 -2.011342  0.178081
9 -0.547744  0.371512  1.231758  0.578528

In [6]: groups = [data[:3], data[3:7], data[7:]]

In [7]: groups
Out[7]:
[          0         1         2         3
 0 -0.150377  0.473775 -0.564428  1.867808
 1  0.178880 -0.356651 -0.864143 -0.325870
 2  0.635222 -0.502338 -0.702672  1.469388,
           0         1         2         3
 3 -0.921042 -0.143417  0.629042 -1.312538
 4  0.224232  0.414366 -0.575869 -1.002100
 5 -1.155228 -0.739624  0.131099 -1.037161
 6 -1.782056 -0.316029 -0.005173  1.159687,
           0         1         2         3
 7  0.878195  0.436940 -0.048127 -1.952570
 8  1.511242 -0.189323 -2.011342  0.178081
 9 -0.547744  0.371512  1.231758  0.578528]

In [8]: pd.concat(groups)
Out[8]:
          0         1         2         3
0 -0.150377  0.473775 -0.564428  1.867808
1  0.178880 -0.356651 -0.864143 -0.325870
2  0.635222 -0.502338 -0.702672  1.469388
3 -0.921042 -0.143417  0.629042 -1.312538
4  0.224232  0.414366 -0.575869 -1.002100
5 -1.155228 -0.739624  0.131099 -1.037161
6 -1.782056 -0.316029 -0.005173  1.159687
7  0.878195  0.436940 -0.048127 -1.952570
8  1.511242 -0.189323 -2.011342  0.178081
9 -0.547744  0.371512  1.231758  0.578528
复制代码

Pandas支持类似sql中的join链接：

In [10]: l = pd.DataFrame({"key": ["foo", "bar"], "lv": [1, 2]})
    ...:

In [11]: ri = pd.DataFrame({"key": ["foo", "bar"], "rv": [4, 5]}
    ...: )

In [12]: l
Out[12]:
   key  lv
0  foo   1
1  bar   2

In [13]: ri
Out[13]:
   key  rv
0  foo   4
1  bar   5
In [15]: pd.merge(l, ri, on='key')
Out[15]:
   key  lv  rv
0  foo   1   4
1  bar   2   5
复制代码

给DataFrame追加行：

In [16]: data = pd.DataFrame(np.random.randn(8, 4), columns=['A'
    ...: , 'B', 'C', 'D'])

In [17]: data
Out[17]:
          A         B         C         D
0 -2.403072  0.523013 -1.730440 -1.050905
1  0.110529  1.797760  0.583266 -0.191529
2  0.308775 -0.904275  0.034278 -1.340783
3  0.931248  0.040340  0.540556 -0.294532
4  0.343270  0.527614 -1.213862 -0.435943
5 -0.887317 -1.292721  0.433839  0.401957
6 -0.037427  0.148965  0.818236 -0.062046
7 -0.537390  0.703600  0.470049  0.420687