Python 如何在Groupby中仅显示具有值的列

你好，数据科学家和熊猫专家

我需要一些帮助，因为我无法正确组织我的数据。 这是我的数据框：

df_dict = [ {'Date': Timestamp('2014-01-03 00:00:00'), 'Store': 'store1', 'employee': 'emp1', 'duties': 'opening'}, \
            {'Date': Timestamp('2014-01-03 00:00:00'), 'Store': 'store1', 'employee': 'emp2', 'duties': 'deli'}, \
            {'Date': Timestamp('2014-01-03 00:00:00'), 'Store': 'store1', 'employee': 'emp3', 'duties': 'cashier'},\
            {'Date': Timestamp('2014-01-03 00:00:00'), 'Store': 'store1', 'employee': 'emp2', 'duties': 'closing'},\
            {'Date': Timestamp('2014-01-03 00:00:00'), 'Store': 'store2', 'employee': 'emp1', 'duties': 'closing'},\
            {'Date': Timestamp('2014-01-03 00:00:00'), 'Store': 'store2', 'employee': 'emp4', 'duties': 'opening'},\
            {'Date': Timestamp('2014-01-03 00:00:00'), 'Store': 'store2', 'employee': 'emp4', 'duties': 'cashier'},\
            {'Date': Timestamp('2014-01-03 00:00:00'), 'Store': 'store2', 'employee': 'emp5', 'duties': 'deli'},\
            {'Date': Timestamp('2014-01-03 00:00:00'), 'Store': 'store3', 'employee': 'emp2', 'duties': 'closing'},\
            {'Date': Timestamp('2014-01-03 00:00:00'), 'Store': 'store3', 'employee': 'emp6', 'duties': 'opening'},\
            {'Date': Timestamp('2014-01-03 00:00:00'), 'Store': 'store3', 'employee': 'emp7', 'duties': 'cashier'},\
            {'Date': Timestamp('2014-01-03 00:00:00'), 'Store': 'store3', 'employee': 'emp6', 'duties': 'deli'},\
            {'Date': Timestamp('2014-01-04 00:00:00'), 'Store': 'store1', 'employee': 'emp1', 'duties': 'opening'},\
            {'Date': Timestamp('2014-01-04 00:00:00'), 'Store': 'store1', 'employee': 'emp2', 'duties': 'deli'},\
            {'Date': Timestamp('2014-01-04 00:00:00'), 'Store': 'store1', 'employee': 'emp3', 'duties': 'cashier'},\
            {'Date': Timestamp('2014-01-04 00:00:00'), 'Store': 'store1', 'employee': 'emp2', 'duties': 'closing'},\
            {'Date': Timestamp('2014-01-04 00:00:00'), 'Store': 'store2', 'employee': 'emp1', 'duties': 'closing'},\
            {'Date': Timestamp('2014-01-04 00:00:00'), 'Store': 'store2', 'employee': 'emp4', 'duties': 'opening'},\
            {'Date': Timestamp('2014-01-04 00:00:00'), 'Store': 'store2', 'employee': 'emp4', 'duties': 'cashier'},\
            {'Date': Timestamp('2014-01-04 00:00:00'), 'Store': 'store2', 'employee': 'emp5', 'duties': 'deli'},\
            {'Date': Timestamp('2014-01-04 00:00:00'), 'Store': 'store3', 'employee': 'emp2', 'duties': 'closing'},\
            {'Date': Timestamp('2014-01-04 00:00:00'), 'Store': 'store3', 'employee': 'emp6', 'duties': 'opening'},\
            {'Date': Timestamp('2014-01-04 00:00:00'), 'Store': 'store3', 'employee': 'emp7', 'duties': 'cashier'},\
            {'Date': Timestamp('2014-01-04 00:00:00'), 'Store': 'store3', 'employee': 'emp6', 'duties': 'deli'},\
            {'Date': Timestamp('2014-01-10 00:00:00'), 'Store': 'store1', 'employee': 'emp1', 'duties': 'opening'},\
            {'Date': Timestamp('2014-01-10 00:00:00'), 'Store': 'store1', 'employee': 'emp2', 'duties': 'deli'},\
            {'Date': Timestamp('2014-01-10 00:00:00'), 'Store': 'store1', 'employee': 'emp3', 'duties': 'cashier'},\
            {'Date': Timestamp('2014-01-10 00:00:00'), 'Store': 'store1', 'employee': 'emp2', 'duties': 'closing'},\
            {'Date': Timestamp('2014-01-10 00:00:00'), 'Store': 'store2', 'employee': 'emp1', 'duties': 'closing'},\
            {'Date': Timestamp('2014-01-10 00:00:00'), 'Store': 'store2', 'employee': 'emp4', 'duties': 'opening'},\
            {'Date': Timestamp('2014-01-10 00:00:00'), 'Store': 'store2', 'employee': 'emp4', 'duties': 'cashier'},\
            {'Date': Timestamp('2014-01-10 00:00:00'), 'Store': 'store2', 'employee': 'emp5', 'duties': 'deli'},\
            {'Date': Timestamp('2014-01-10 00:00:00'), 'Store': 'store3', 'employee': 'emp2', 'duties': 'closing'},\
            {'Date': Timestamp('2014-01-10 00:00:00'), 'Store': 'store3', 'employee': 'emp6', 'duties': 'opening'},\
            {'Date': Timestamp('2014-01-10 00:00:00'), 'Store': 'store3', 'employee': 'emp7', 'duties': 'cashier'},\
            {'Date': Timestamp('2014-01-10 00:00:00'), 'Store': 'store3', 'employee': 'emp6', 'duties': 'deli'}]

我想按如下方式组织我的输出：

                     Store 1               Store 2          store3      
    Week          emp1  emp2  emp3     emp1 emp4 emp5   emp2 emp6 emp7
    2013-12-30     2    4       2        2    4   2      2    4    2
    2014-01-06     1    1       1        1    1   1      2    1    1

因此，我试着按表达式分组如下：

df_group = dict_df.groupby([pd.Grouper(key='Date', freq='W-MON'), 'Store', 'employee'])\
                            ['duties'].count().unstack(level=1).unstack(level=1).reset_index()

但是，它显示的是所有员工，而不是该特定商店中的员工工作示例：

                      Store 1                            
Week          emp1  emp2  emp3 emp4 emp5 emp6  emp7 
2013-12-30     2    4       2   NaN NaN  NaN   NaN 
2014-01-06     1    1       1   NaN NaN  NaN   NaN

那么我怎样才能得到我想要的结果呢。基本上我想过滤掉那些不在那家商店工作的员工

最好是使用GROMPBY来满足这个需求，还是应该考虑其他方法？< /P>

---

提前感谢您的帮助和考虑。

尝试取消堆叠多个级别

[1,2]

：

df_out = (df.groupby([pd.Grouper(key='Date', freq='W-MON'), 'Store', 'employee'])['duties']
            .count()
            .unstack(level=[1, 2])
        )
print(df_out)

印刷品：

Store      store1           store2           store3          
employee     emp1 emp2 emp3   emp1 emp4 emp5   emp2 emp6 emp7
Date                                                         
2014-01-06      2    4    2      2    4    2      2    4    2
2014-01-13      1    2    1      1    2    1      1    2    1

---

可以同时取消堆叠两个标高：

(df.groupby([pd.Grouper(key='Date', freq='W-MON'), 'Store','employee'])
   .size().unstack(['Store','employee'])
)

输出：

Store      store1           store2           store3          
employee     emp1 emp2 emp3   emp1 emp4 emp5   emp2 emp6 emp7
Date                                                         
2014-01-06      2    4    2      2    4    2      2    4    2
2014-01-13      1    2    1      1    2    1      1    2    1

---

很好的解决方案，我不知道你可以按名称堆叠/取消堆叠：）谢谢你，Quang，这很有效。非常感谢您的及时回复。事实上，我学到了一个新的东西，我们可以使用列名称来使用unstack。嗨，Andrej，谢谢你们的及时回复。我也试过了，效果不错。但我喜欢指定列名，因为代码看起来更可读。非常感谢您的及时回复。