Question

我的df是这样的：

year    state     party         percentage 
1976    Arizona   republican    43.34
1976    Arizona   third party   0.21
1976    Arizona   democrat      54.01
1976    Arizona   third party   0.99
1976    Arizona   third party   .45
1978    Alabama   third party   6.01
1978    Alabama   republican    43.32
1978    Alabama   third party   0.82
1978    Alabama   democrat      55.06 
1978    Alabama   democrat      93.99
1978    Alabama   third party   0.80

我想使用.groupby，但如果一方是“第三方”，则仅求和。这是我的代码：

g = df_senate.groupby(['year','state','party'], as_index=False)
g.apply(lambda x: x[x['party'] == 'third party']['percentage'].sum())

几乎可以使用，但是其他“一方”的值却为0，并且汇总了其他各方。我想不对每年/州的其他“民主”和“共和党”行进行汇总。我只想总结“第三方”：

year  state       party          percentage
1976  Arizona     democrat       0.00
                  republican     0.00
                  third party    2.65
1978  Alabama     democrat       0.00
                  republican     0.00
                  third party    7.63

我又如何将其保留为数据框？将as_index=False放入.groupby无效。我要结束的是：

year  state       party          percentage
1976    Arizona   republican    43.34
1976    Arizona   third party   2.65
1976    Arizona   democrat      54.01
1978    Alabama   third party   7.63
1978    Alabama   republican    43.32
1978    Alabama   democrat      55.06 
1978    Alabama   democrat      93.99

（如果您想知道，这是参议院的选举数据，有时一个州将不得不选举2名参议员，而不仅仅是一个参议员，而且我的选民总数不必超过100％，因为那太奇怪了）

提前谢谢！

Answer 1

df = (
        df.groupby(['year', 'state', 'party'])['percentage'].apply(
            lambda x: [x.sum()] if x.name[2] == 'third party' else list(x))
        .explode()
        .reset_index() 
      )

print(df)

打印：

   year    state        party percentage
0  1976  Arizona     democrat      54.01
1  1976  Arizona   republican      43.34
2  1976  Arizona  third party       1.65
3  1978  Alabama     democrat      55.06
4  1978  Alabama     democrat      93.99
5  1978  Alabama   republican      43.32
6  1978  Alabama  third party       7.63

Python Pandas条件和，同时保留其他值

1 个答案: