以另外两列pandas为条件创建一个新列

Question

I have a dataframe with two columns. I want to create a new column and input whichever column has the longest string. so

        column_a        column_b             column_c

   0  'dog is fast'   'dog is faster'      'dog is faster' (desired output)

I tried this code but got an error saying that int is not iterable, I was thinking in merging the series after to the df. I wasn't sure how to implement it right away into a column of the df.

column_c = pd.Series()

 for i in len(df.column_a):
  if len(df.column_a.iloc[i]) >= len(df.column_b.iloc[0]):
    column_c.append(df.column_a.iloc[i])
  else:
    column_c.append(df.column_b.iloc[i])

any help is apreciated.

Answer 1

Use pandas.DataFrame.apply :

Given sample data

import pandas as pd

df = pd.DataFrame([['fast', 'faster'], ['slower', 'slow']])
        0       1
0    fast  faster
1  slower    slow

df['column_c'] = df.apply(lambda x:max(x, key=len), 1)

Output:

        0       1 column_c
0    fast  faster   faster
1  slower    slow   slower

Answer 2

Using np.where with str.len

df['column_c']=np.where(df.column_a.str.len()>df.column_b.str.len(),df.column_a,df.column_b)
df
Out[301]: 
        column_a         column_b         column_c
0  'dog is fast'  'dog is faster'  'dog is faster'

Answer 3

可以使用 df.apply()

df['column_c'] = df.apply(lambda x: x[0] if len(x[0]) > len(x[1]) else x[1], axis=1)

Answer 4

You can use DataFrame.apply . You need to apply on specific columns if you have more than two columns in your dataframe

df['column_c'] = df.apply(lambda x: x[0] if len(x[0]) > len(x[1]) else x[1], axis = 1)

     column_a        column_b        column_c
0   'dog is fast'   'dog is faster' 'dog is faster'