Pandas pivot 還是 pivot 表？

Question

我正在嘗試使用以下格式獲取 excel 數據框：

    DOW                 Location    7/30/2022   8/6/2022    8/13/2022   8/20/2022
    Volumes Saturday    North       33          32          29          24
    Volumes Saturday    South       34          17          30          28
    Volumes Sunday      North       40          57          25          28
    Volumes Sunday      South       47          38          32          45
    ACT Saturday        North       750         1060        1066        1082
    ACT Saturday        South       545         509         1306        1121
    ACT Sunday          North       801         860         572         795
    ACT Sunday          South       622         526         711         491

並以這種格式擁有 output：

DOW_Location            Location    Date        Volumes  ACT
Volumes Saturday North  North       7/30/2022   33       750
Volumes Sunday North    North       7/31/2022   40       801                
Volumes Saturday South  South       7/30/2022   34       545
Volumes Sunday South    South       7/31/2022   47       622

我想我可以在通過組合df['DOW'] + df['Location']創建唯一索引后執行df.pivot()並從df['DOW']中刪除工作日以創建df['Category']列。 我的問題是，當值跨越多個列時，我不知道如何填充 pivot 中的值。

所以我不確定這里的“值”參數是什么：

df = df.pivot(index='DOW_Location', columns='Category', values=?)

如果 pivot 不是這里的理想解決方案，那么另一種方法是什么？

Answer 1

# split the DOW into columns
df[['col','day']]=df['DOW'].str.split(' ',expand=True)


# melt to bring the dates as rows
df2=df.melt(id_vars=['DOW','Location','col','day'], var_name='date')


#pivot and sum
(df2.pivot_table(index=['day','Location','date'],
                columns=['col'],aggfunc=sum)
    .reset_index())

    day     Location    date    value
col                 ACT     Volumes
0   Saturday    North   7/30/2022   750     33
1   Saturday    North   8/13/2022   1066    29
2   Saturday    North   8/20/2022   1082    24
3   Saturday    North   8/6/2022    1060    32
4   Saturday    South   7/30/2022   545     34
5   Saturday    South   8/13/2022   1306    30
6   Saturday    South   8/20/2022   1121    28
7   Saturday    South   8/6/2022    509     17
8   Sunday  North   7/30/2022   801     40
9   Sunday  North   8/13/2022   572     25
10  Sunday  North   8/20/2022   795     28
11  Sunday  North   8/6/2022    860     57
12  Sunday  South   7/30/2022   622     47
13  Sunday  South   8/13/2022   711     32
14  Sunday  South   8/20/2022   491     45
15  Sunday  South   8/6/2022    526     38