python - Pandas 分组/透视数据，而一列的条目成为新标签

Question

我想用python + pandas按技术总结发电厂的容量( previous question )。

对于此任务，数据必须进行分组/透视，而“技术”列中的列条目应成为列标签

这是我的输入:

Plant Name,Nameplate Capacity,Technology,...
Barry,153.1,Natural Gas Steam Turbine,..
Barry,153.1,Natural Gas Steam Turbine,..
Barry,403.7,Conventional Steam Coal,..
Barry,788.8,Conventional Steam Coal,..
Barry,195.2,Natural Gas Fired Combined Cycle,..
Barry,195.2,Natural Gas Fired Combined Cycle,..

以及所需的输出:

Plant Name,Natural Gas Steam Turbine,Conventional Steam Coal,Natural Gas Fired Combined Cycle,..
Barry,306.2,1192.5,390.4,..

我尝试了一些命令，但没有成功:

df.groupby(['Plant Name', 'Technology']).sum().pivot('Plant Name', 'Technology').fillna(0)

或

#with numpy as np
res = df.pivot_table(index=["Plant Name"], columns=["Plant Name"], values=["Technology"], aggfunc=np.sum)

一个附加问题

如何找出每行的最大条目(例如我的示例中的“常规动力煤”)作为新列？

Answer 1

我认为需要更改列名称并添加参数fill_value:

res = df.pivot_table(index="Plant Name", 
                     columns="Technology", 
                     values="Nameplate Capacity", 
                     aggfunc=np.sum,
                     fill_value=0).reset_index()
print (res)
Technology Plant Name  Conventional Steam Coal  \
0               Barry                   1192.5   

Technology  Natural Gas Fired Combined Cycle  Natural Gas Steam Turbine  
0                                      390.4                      306.2  

第一个解决方案应更改为聚合 sum 和 unstack 的指定列 reshape :

res = (df.groupby(['Plant Name', 'Technology'])['Nameplate Capacity']
         .sum()
         .unstack(fill_value=0)
         .reset_index())
print (res)
Technology Plant Name  Conventional Steam Coal  \
0               Barry                   1192.5   

Technology  Natural Gas Fired Combined Cycle  Natural Gas Steam Turbine  
0                                      390.4                      306.2