Pivot一个多列的pandas数据框架。

Posted 2021-03-29

我有一个示例数据框架，如下所示

df1 = pd.DataFrame({'Gender':['Male','Male','Male','Male','Female','Female','Female','Female','Male','Male','Male','Male','Female','Female','Female','Female'],
                'Year' :[2008,2008,2009,2009,2008,2008,2009,2009,2008,2008,2009,2009,2008,2008,2009,2009],
           'rate':[2.3,3.2,4.5,6.7,5.6,3.2,3.5,2.6,2.3,3.2,4.5,6.7,5.6,3.2,3.5,2.6],
           'Heading':['TNMAB123','TNMAB123','TNMAB123','TNMAB123','TNMAB123','TNMAB123','TNMAB123','TNMAB123',
                     'TNMAB456','TNMAB456','TNMAB456','TNMAB456','TNMAB456','TNMAB456','TNMAB456','TNMAB456'],
           'target':[31.2,33.4,33.4,35.2,35.2,36.4,36.4,37.2,31.2,33.4,33.4,35.2,35.2,36.4,36.4,37.2],
            'day_type':['wk','wkend','wk','wkend','wk','wkend','wk','wkend','wk','wkend','wk','wkend','wk','wkend','wk','wkend']})

我想把它们进行转置，得到如下图所示的输出，但我的代码却出现了如下图所示的错误。

df1.pivot(index='Year', columns='Heading', values='rate')

在SO贴的帮助下，我写了这个，但是对于3列，我不知道如何使其工作？

df1 = df1.pivot_table(index=['Year','Gender','day_type'],columns='Heading',values='rate').unstack()
df1.columns = ['_'.join(i) for i in df1.columns.tolist()]

我希望我的输出是如下图所示，其中每一年作为一行，该年的所有对应条目作为列。

请注意，我没有填写数值，因为表列结构更重要。

答案

尝试用 map，你还需要 unstack 两种 level

df1 = df1.pivot_table(index=['Year','Gender','day_type'],columns='Heading',values='rate').unstack([1,2])
df1.columns=df1.columns.map('_'.join)
df1
      TNMAB123_Female_wk  ...  TNMAB456_Male_wkend
Year                      ...                     
2008                 5.6  ...                  3.2
2009                 3.5  ...                  6.7
[2 rows x 8 columns]