python - 有没有办法对包含列表的数据框列进行排序？

Question

我在数据框中有一列包含列表数据

例如:

enter code here
| Column 1 | Column 2                  |
|--------- |---------------------------|
|    1     | ['jan']                   |
|    2     | ['jan','feb']             |
|    3     | ['feb','jan']             |
|    4     | ['mar','april','jan']     |
|    5     | ['jan','mar','april']     |
|    6     | ['jan','feb','mar']       |
|    7     | ['jan','feb','mar']       |

如上表所示，我们可以看到第 2 列、第 2 行和第 3 行具有相同的列表，但顺序不同。其他值也是如此。像(第 4 行和第 5 行)

我必须按这个顺序排序

enter code here ['oct' ,'nov' ,'dec' ,'jan' ,'feb' ,'march' ,'april']

所需的输出应如下所示，一个新的第 3 列，其中包含已排序的值

enter code here
| Column 1 | Column 2                  |      Column 3             |
|--------- |---------------------------|---------------------------|
|    1     | ['oct']                   | ['oct']                   |
|    2     | ['oct','nov']             | ['oct','nov']             |
|    3     | ['nov','oct']             | ['oct','nov']             |
|    4     | ['mar','april','jan']     | ['jan','mar','april']     |
|    5     | ['jan','mar','april']     | ['jan','mar','april']     |
|    6     | ['jan','feb','mar']       | ['jan','feb','mar']       |
|    7     | ['jan','feb','mar']       | ['jan','feb','mar']       |

Answer 1

使用 sorted 通过 enumerate 创建字典:

#import ast

#converted to lists if necessary
df['Column 2'] = df['Column 2'].apply(ast.literal_eval)

months = ['oct' ,'nov' ,'dec' ,'jan' ,'feb' ,'mar' ,'april']

d = {v: k for k, v in enumerate(months)}
   
df['Column 2'] = df['Column 2'].apply(lambda x: sorted(x, key=lambda y: d[y]))
print (df)
   Column 1           Column 2
0         1              [jan]
1         2         [jan, feb]
2         3         [jan, feb]
3         4  [jan, mar, april]
4         5  [jan, mar, april]
5         6    [jan, feb, mar]
6         7    [jan, feb, mar]

如果值是前 3 个字母，则可以按日期时间排序:

df['Column 2'] = df['Column 2'].apply(lambda x: sorted(x, key=lambda y: pd.to_datetime(y, format='%b')))
print (df)
   Column 1         Column 2
0         1            [jan]
1         2       [jan, feb]
2         3       [jan, feb]
3         4  [jan, mar, apr]
4         5  [jan, mar, apr]
5         6  [jan, feb, mar]
6         7  [jan, feb, mar]