将 Pandas df 单元格转换为列名

Question

I have a df like this:

0   1   2   3   4   5   
abc 0   1   0   0   1
bcd 0   0   1   0   0   
def 0   0   0   1   0

How can I convert the dataframe cells to be the column name if there's a 1 in the cell? Looks like this:

0   1   2   3   4   5   
abc 0   2   0   0   5
bcd 0   0   3   0   0   
def 0   0   0   4   0

Answer 1

Let us try

df.loc[:,'1':] = df.loc[:,'1':]  * df.columns[1:].astype(int)
df
Out[468]: 
     0  1  2  3  4  5
0  abc  0  2  0  0  5
1  bcd  0  0  3  0  0
2  def  0  0  0  4  0

Answer 2

I'd suggest you simply do the logic for each column, where the value is 1 in the given column, set the value as the column name

for col in df.columns:
    df.loc[df[col] == 1, col] = col

Answer 3

We can use np.where over the whole dataframe:

values = np.where(df.eq(1), df.columns, df)
df = pd.DataFrame(values, columns=df.columns)

     0  1  2  3  4  5
0  abc  0  2  0  0  5
1  bcd  0  0  3  0  0
2  def  0  0  0  4  0