Efficient way to replace values in one column using another column in pandas

Question

How do I replace string values in a dataframe column[1] using list of string values in a different column[2].

Data

          0                       1            2              3
0  3000 20%  dummy1 3000 dummy2 20%  [3000, 20%]  dummy1 dummy2

I want to replace string value in column 1 i.e. "dummy1 3000 dummy2 20%" using list in column 2 i.e. "[3000, 20%]". So 3000 and 20% are replaced with ""(empty string) from the string to form 3rd column(Result) i.e. "dummy1 dummy2"

Code

df = pd.DataFrame([['3000 20%', 'dummy1 3000 dummy2 20%']])
df[2] = df[0].str.split(' ')

def replace_string(x):
    repl_string = str(x[1])
    for key in x[2]:
        repl_string = repl_string.replace(key, '')
    return ' '.join(repl_string.split())

df[3] = df.apply(replace_string, axis=1)

I have currently written the above code, which is slow for large dataframe. How do I improve the efficiency of this code or is there any other way to do this?

Please add more information to this question. Do you want all all values in column 1 to be replaced with the values of column 2? While this can easily be done using df[1] = df[2].apply(lambda x: x), it does not make any sense to do, so the question is are you trying to replace null values or what? — Samuel Nde
– Samuel Nde, Commented Oct 4, 2018 at 5:33
I have added more info on what to replace with what. If needed any more please comment here. — Vishnudev Krishnadas
– Vishnudev Krishnadas, Commented Oct 4, 2018 at 5:50

jezrael · Accepted Answer · 2018-10-04 05:51:08Z

1

Use nested list comprehension:

df = pd.DataFrame([['3000 20%', 'dummy1 a 3000 dummy2 20%'],
                   ['abc 2%', 'klmn 3000 dummy2 2%']])
print (df)
          0                         1
0  3000 20%  dummy1 a 3000 dummy2 20%
1    abc 2%       klmn 3000 dummy2 2%

df[3] = [' '.join(y for y in j.split() if y not in i.split()) for i, j in zip(df[0], df[1])]
print (df)
          0                         1                 3
0  3000 20%  dummy1 a 3000 dummy2 20%   dummy1 a dummy2
1    abc 2%       klmn 3000 dummy2 2%  klmn 3000 dummy2

answered Oct 4, 2018 at 5:51

jezrael

868k103 gold badges1.4k silver badges1.3k bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Efficient way to replace values in one column using another column in pandas

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related