Question

我有一个如下所示的DataFrame df：

|  A    |  B  | ... |
---------------------
| one   | ... | ... |
| one   | ... | ... |
| one   | ... | ... |
| two   | ... | ... |
| three | ... | ... |
| three | ... | ... |
| four  | ... | ... |
| five  | ... | ... |
| five  | ... | ... |

正如您在A所看到的，有5个唯一值。我想随机拆分DataFrame。例如，我希望DataFrame df1中有3个唯一值，DataFrame df2中有2个唯一值。我的问题是它们不是唯一的。我不想在两个DataFrame上拆分这些唯一值。

因此生成的DataFrame可能如下所示：

DataFrame df1，包含3个唯一值：

|  A    |  B  | ... |
---------------------
| one   | ... | ... |
| one   | ... | ... |
| one   | ... | ... |
| three | ... | ... |
| three | ... | ... |
| five  | ... | ... |
| five  | ... | ... |

DataFrame df2，包含2个唯一值：

|  A    |  B  | ... |
---------------------
| two   | ... | ... |
| four  | ... | ... |

无论如何如何轻松实现这一目标？我想过分组，但我不确定如何从这个分裂...

Answer 1

<强>设置

df=pd.DataFrame({'A': {0: 'one',
  1: 'one',
  2: 'one',
  3: 'two',
  4: 'three',
  5: 'three',
  6: 'four',
  7: 'five',
  8: 'five'},
 'B': {0: 0, 1: 1, 2: 2, 3: 3, 4: 4, 5: 5, 6: 6, 7: 7, 8: 8}})

<强>解决方案

#get 2 unique keys from column A for df1. You can control the split either
# by absolute number in each group, or by a percentage. Check docs for the .sample() func.
df1_keys = df.A.drop_duplicates().sample(2)
df1 = df[df.A.isin(df1_keys)]
#anything not in df1_keys will be assigned to df2
df2 = df[~df.A.isin(df1_keys)]

df1_keys
Out[294]: 
7    five
0     one
Name: A, dtype: object

df1
Out[295]: 
      A  B
0   one  0
1   one  1
2   one  2
7  five  7
8  five  8

df2
Out[296]: 
       A  B
3    two  3
4  three  4
5  three  5
6   four  6

Answer 2

.isin()

最后，使用r1 = df[df['A'].isin(v1)] r2 = df[df['A'].isin(v2)]方法索引数据框以获得所需的结果。

Optional

随机拆分DataFrame（取决于唯一值）

2 个答案: