How to collapse Pandas Dataframe Columns and Concatenate Strings

问题

I have a Data Frame df0 with n columns. Only one of the columns contains a string, all other columns are empty or contain the "" string.

Is it possible to collapse the data frame into a single column data frame where for each row I get the non-empty element?

df0:

    A    B     C
1  Car
2  Car 
3       Bike
4  Car
5            Train
6            Train

should give:

    1    
1  Car
2  Car 
3  Bike
4  Car
5  Train
6  Train

回答1:

Maybe:

>>> df.max(axis=1)
1      Car
2      Car
3     Bike
4      Car
5    Train
6    Train
dtype: object

which is a Series, not a DataFrame, but you could make one using df.max(axis=1).to_frame(1) or something.

回答2:

If they are empty strings rather than NaN you can use .sum:

In [11]: df.fillna('').sum(1)
Out[11]: 
1      Car
2      Car
3     Bike
4      Car
5    Train
6    Train
dtype: object

回答3:

This would also work, under the assumption there's always a column with a non empty string:

 df.apply(lambda x: [y for y in x.values if y!=''][0],axis=1)

来源：https://stackoverflow.com/questions/22491845/how-to-collapse-pandas-dataframe-columns-and-concatenate-strings