如何在python中转换为整齐的格式？

Question

My pandas dataframe has separate columns that are one-hot encoded and a total column at the end that sums them up ( total = val1 + val2 ).

Some rows have 1s for multiple val columns:

| name | val1 | val2 | total | 
| joe  | 1    | 0    | 1     |
| bob  | 0    | 1    | 1     |
| dan  | 1    | 1    | 2     |

I want this:

| name | val1 | val2 | total | 
| joe  | 1    | 0    | 1     |
| bob  | 0    | 1    | 1     |
| dan  | 1    | 0    | 1     |
| dan  | 0    | 1    | 1     |

I can't figure out how to get this to work: to melt it conditional upon the total column.

The end result should have a total value of 1 for every row.

Answer 1

d = df.drop('total', axis=1).set_index('name').stack().loc[lambda x: x == 1]

n, v = zip(*d.index)
pd.concat([pd.Series(n, name='name'), pd.get_dummies(v).assign(total=1)], axis=1)

  name  val1  val2  total
0  joe     1     0      1
1  bob     0     1      1
2  dan     1     0      1
3  dan     0     1      1

Answer 2

Harder than what I thought

s1=df.iloc[:,1:-1]
s2=df.iloc[:,0]
df[['name']].join(s1.mul(s2,0).replace('',np.nan).stack().reset_index(level=1)['level_1'].str.get_dummies(),how='right').assign(Total=1)
Out[413]: 
  name  val1  val2  Total
0  joe     1     0      1
1  bob     0     1      1
2  dan     1     0      1
2  dan     0     1      1