CODEWITHSUNDEEP
pythonpandasdataframe
LearnerBegineer
LearnerBegineer

Reputation: 23

How to resolve TypeError: string indices must be integers, while applying lambda function in pandas?

I am applying a lambda function on my DataFrame. But I think I am doing it wrong. What is the correct way to apply lambda function ?

input:

ID Name result old_result

A1 Jim Bad Good

A2 Tim Good Good

A3 Matt Good None

code:

df3['avg_result'] = df3['result'].apply(lambda x : x['old_result'] if x['result'] == 'Bad' else x['result'])

Expected Output:

ID Name result old_result avg_result

A1 Jim Bad Good Good

A2 Tim Good Good Good

A3 Matt Good None Good

Upvotes: 0

Views: 844

Answers (1)

Shubham Periwal
Shubham Periwal

Reputation: 2248

You are doing df['result'].apply(lambda x ...) so the value of x actually is each individual string in the column df['result']. That's why when you try to do x['old_result'], it says "string indices must be integers" because it is doing something like Good['old_result'] which is not possible.

What you need is

df['avg_result'] = df.apply(lambda x : x['old_result'] if x['result'] == 'Bad' else x['result'], axis=1)

What this does is instead of applying the lambda function to each string in df['result'] column, it applies the function to each row in the dataframe (that's where axis=1 comes in).

And within that row you can then do x['old_result'] and it will return you the value of old_result column in that row.

Output

   ID   Name    result  old_result  avg_result
0   A1  Jim     Bad     Good    Good
1   A2  Tim     Good    Good    Good
2   A3  Matt    Good    None    Good

Recommendation

In fact, you can do the same thing in a much more readable manner with np.where instead of using .apply and lambda

df['npwhere_result'] = np.where(df['result']=='Bad', df['old_result'], df['result'])

    ID  Name    result  old_result  avg_result  npwhere_result
0   A1  Jim     Bad     Good         Good       Good
1   A2  Tim     Good    Good         Good       Good
2   A3  Matt    Good    None         Good       Good

Upvotes: 2

Related Questions