熊猫数据框按条件拆分

Question

我是 pandas 的新手，如果我听起来太天真，请原谅。我有两个数据框 df1 和 df2，

df1 = pd.DataFrame({'key1': ['K0', 'K1', 'K2', 'K3'],
               'key2': ['K5', 'K4', 'K5', 'K4']})

df2 = pd.DataFrame({'key1': ['K0', 'K1', 'K2', 'K3', 'K9', 'K8', 'K7'],
                   'key2': ['K5', 'K6', 'K5', 'K4', 'K6', 'K4', 'K5'],
                     'A':['1', '2', '3', '4', '5', '6', '7'],
                     'B':['8', '9', '10', '11', '12', '13', '14']})

我想将 df2 合并到 df1 上

final = df1.merge(df2, on=['key1', 'key2'], how='left')

然后将 df2 中的剩余值作为一个数据帧。

任何帮助将不胜感激。谢谢。

Answer 1

IIUC 你需要外连接参数indicator，然后用boolean indexing分割：

final = df1.merge(df2, how='outer', indicator=True)
print (final)
  key1 key2    A    B      _merge
0   K0   K5    1    8        both
1   K1   K4  NaN  NaN   left_only
2   K2   K5    3   10        both
3   K3   K4    4   11        both
4   K1   K6    2    9  right_only
5   K9   K6    5   12  right_only
6   K8   K4    6   13  right_only
7   K7   K5    7   14  right_only

print (final[final._merge == 'right_only'])
  key1 key2  A   B      _merge
4   K1   K6  2   9  right_only
5   K9   K6  5  12  right_only
6   K8   K4  6  13  right_only
7   K7   K5  7  14  right_only

print (final[final._merge != 'right_only'])
  key1 key2    A    B     _merge
0   K0   K5    1    8       both
1   K1   K4  NaN  NaN  left_only
2   K2   K5    3   10       both
3   K3   K4    4   11       both

---

print (final[final._merge == 'right_only'].drop('_merge', axis=1))
  key1 key2  A   B
4   K1   K6  2   9
5   K9   K6  5  12
6   K8   K4  6  13
7   K7   K5  7  14

print (final[final._merge != 'right_only'].drop('_merge', axis=1))
  key1 key2    A    B
0   K0   K5    1    8
1   K1   K4  NaN  NaN
2   K2   K5    3   10
3   K3   K4    4   11