1

So, I have a function like below:

def do_something(row, args):
    # doing something
    return arr

where arr = array([1, 2, 3, 4, 5])

and in my main function, I have a pandas DataFrame object df, it has columns A, B, C, D, E along with other columns.

I use function do_something as below, try to assign each of the elements of the returning array to the columns:

df[['A', 'B', 'C', 'D', 'E']] = df.apply(lambda row: do_something(row, args), axis=1)

However, it gave me an error:

ValueError: Must have equal len keys and values when setting with an iterable

I think it means I'm trying to assign the whole array as one variable to multiple columns in my df, hence the unequal length trigger the error. But I'm not sure how to make my goal.

Improve this question
2
  • What is the size of your df? Commented Mar 9, 2022 at 21:46
  • @ZLi 10 columns and 5k rows. Commented Mar 9, 2022 at 21:47

3 Answers 3

Reset to default
1

The error is telling you that you are trying to assign a sequence of values to a sequence of keys. You can use the zip function to create a dictionary from the two lists. Then you can use the dictionary to assign the values to the columns.

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

1 Comment

how to make it when using the apply function?
0

This is because the return value of apply is a pd.Series which comes with 1 column while you try to assign it to 5. See this example:

import numpy as np
import pandas as pd

def do_something(row, args):
    arr = np.array([1,2,3,4,5])
    return arr 
df = pd.DataFrame(np.ones((2,6)),columns = ['A', 'B', 'C', 'D', 'E', 'some other column'])
df.apply(lambda row: do_something(row, 2), axis=1)

0    [1, 2, 3, 4, 5]
1    [1, 2, 3, 4, 5]
dtype: object

The solution is to convert it to 5 columns so you could do the assign later:

df[['A', 'B', 'C', 'D', 'E']] = df.apply(lambda row: do_something(row, 2), axis=1).to_list()
Improve this answer

3 Comments

Thanks for your thoughts. However, I need the other columns in my df to carry out some calculation and your solution means I need to create a new df which only have the 5 columns. I tried the tolist() method though, but it gave me a new error: ValueError: cannot reindex from a duplicate axis. I checked my df's index and they are all unique, so is the array as it is just one row.
I don't get your question. The method above does allow additional columns in the calculation. It just returns 5 columns (or any number depends on your function). As for your error I would guess your returned value contains its own index. You could print out the returned value of your apply function and check
@Z Li I see what you mean. I'm actually using the same method as your solution that expand the df with columns ABCDE, then run the apply with tolist(), but it gave me the error as mentioned in my last comment.
0

you can use the parameter "result_type" of the function apply. Setting this parameter to "expand" will turn the resulting list into columns.

df[['A', 'B', 'C', 'D', 'E']] = df.apply(lambda row: do_something(row, args), result_type = 'expand', axis=1)

Check this

Improve this answer

Comments

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.