具有不均匀组的数据透视表 Python Pandas

Question

我有一个 table，我想使用 Python 3x Pandas 将其转换为以下内容：

Group	Assessment	Review
GroupA	No team spirit	Negative
GroupA	Good players	Positive
GroupA	They scored well	Positive
GroupB	Goal failed	Negative
GroupB	Bad weather	Negative
GroupB	Resilience	Positive
GroupB	Growth potential	Positive
GroupB	Bad technique	Negative

结果 table 应该是：

Group	Positive	Negative
GroupA	Good players	No team spirit
GroupA	They scored well	NaN
GroupB	Resilience	Goal Failed
GroupB	Growth Potential	Bad weather
GroupB	NaN	Bad Technique

是否有使用 Pandas 或其他方法的简洁 Pythonic 方法？

Answer 1

您不能按原样 pivot，但您可以使用 groupby + cumcount 和 pivot 添加组号，使用新创建的编号作为索引：

out = (df.assign(num=df.groupby(['Group','Review']).cumcount())
       .pivot(['num','Group'],'Review','Assessment')
       .droplevel(0).sort_index()
       [['Positive','Negative']]
       .reset_index()
       .rename_axis(columns=[None]))

输出：

    Group          Positive        Negative
0  GroupA      Good players  No team spirit
1  GroupA  They scored well             NaN
2  GroupB        Resilience     Goal failed
3  GroupB  Growth potential     Bad weather
4  GroupB               NaN   Bad technique

Answer 2

这不是您真正想要的，但这个 table 结构似乎更有意义：

res = (df.groupby(['Group','Review']).
       apply(lambda x:x['Assessment'].tolist()).
       unstack())
print(res)
'''
Review                                   Negative                          Positive
Group                                                                              
GroupA                           [No team spirit]  [Good players, They scored well]
GroupB  [Goal failed, Bad weather, Bad technique]    [Resilience, Growth potential]

具有不均匀组的数据透视表 Python Pandas

Pivot Data with Uneven Groups Python Pandas

python

pivot

dataframe

python-3.x

pandas