填写丢失的数据熊猫

Question

如何在此日期框架中填写缺失的数据。

在没有销售的情况下缺少数天的值。如何在特定商店和日期销售0件商品的日子里填写缺失值？

输入

Dates            Store            Item        Sales           
2017-01-01       Chicago          Apple       10
2017-01-02       NewYork          Pear        10 
2017-01-03       Chicago          Apple       10

输出

Dates            Store            Item        Sales           
2017-01-01       Chicago          Apple       10
2017-01-01       Chicago          Pear        0
2017-01-02       Chicago          Apple       0
2017-01-02       Chicago          Pear        0
2017-01-03       Chicago          Apple       10
2017-01-03       Chicago          Pear        0    
2017-01-01       NewYork          Apple       0
2017-01-01       NewYork          Pear        0 
2017-01-02       NewYork          Apple       0 
2017-01-02       NewYork          Pear        10 
2017-01-03       NewYork          Apple       0 
2017-01-03       NewYork          Pear        0 

Answer 1

使用：

• Multiindex
的第一个set_index
• 制作新的Multiindex from_product
• reindex并添加0以查找缺失值
• sort_index和reset_index
的最后排序级别Store

df = df.set_index(['Dates','Store','Item'])
mux = pd.MultiIndex.from_product(df.index.levels, names=df.index.names)
df = df.reindex(mux, fill_value=0).sort_index(level='Store').reset_index()
print (df)
        Dates    Store   Item  Sales
0  2017-01-01  Chicago  Apple     10
1  2017-01-01  Chicago   Pear      0
2  2017-01-02  Chicago  Apple      0
3  2017-01-02  Chicago   Pear      0
4  2017-01-03  Chicago  Apple     10
5  2017-01-03  Chicago   Pear      0
6  2017-01-01  NewYork  Apple      0
7  2017-01-01  NewYork   Pear      0
8  2017-01-02  NewYork  Apple      0
9  2017-01-02  NewYork   Pear     10
10 2017-01-03  NewYork  Apple      0
11 2017-01-03  NewYork   Pear      0

Answer 2

使用set_index，stack和unstack

会有点难以理解

df.set_index(['Dates','Store','Item']).unstack().stack(dropna=False).\
    unstack(1).stack(dropna=False).fillna(0).reset_index()
Out[258]: 
         Dates   Item    Store  Sales
0   2017-01-01  Apple  Chicago   10.0
1   2017-01-01  Apple  NewYork    0.0
2   2017-01-01   Pear  Chicago    0.0
3   2017-01-01   Pear  NewYork    0.0
4   2017-01-02  Apple  Chicago    0.0
5   2017-01-02  Apple  NewYork    0.0
6   2017-01-02   Pear  Chicago    0.0
7   2017-01-02   Pear  NewYork   10.0
8   2017-01-03  Apple  Chicago   10.0
9   2017-01-03  Apple  NewYork    0.0
10  2017-01-03   Pear  Chicago    0.0
11  2017-01-03   Pear  NewYork    0.0