How to perform string replace operation using pandas in python?

Question

I have a pandas dataframe where the Items columns is a list.

cust_id   Items
100     ['item1','item2','item3']
101     ['item5','item8','item9']
102     ['item2','item4']

I want to convert the above dataframe to the below format.

cust_id  Items
100     item1 item2 item3
101     item5 item8 item9
102     item2 item4

I tried using the pandas built in string replace function put its returning the original column without actually performing the string replace operation.

df['Items']=(df['Items'].astype(str)).replace({"['":"", "', '":" ", "']":"" },method='string')

Please advise

Update:

I used the below code to create the original dataframe.

df=df1.groupby(['cust_id'])['Items'].apply(list).reset_index()

Sorry are the elements lists or a string of a list? Can you post code to construct your df to avoid ambiguity — EdChum
– EdChum, Commented Sep 23, 2015 at 12:08
Why not just use an apply on the column, and do something like lambda lst: ' '.join(lst) — Brian Pendleton
– Brian Pendleton, Commented Sep 23, 2015 at 12:10
@EdChum I have added the code to reconstruct my original df. — GeorgeOfTheRF
– GeorgeOfTheRF, Commented Sep 23, 2015 at 13:08

Anand S Kumar · Accepted Answer · 2015-09-23 13:16:24Z

4

If the elements are really list , then you can use str.join() on the list along with series.apply method . Example -

In [159]: df = pd.DataFrame([[100,['item1','item2','item3']],[101,['item5','item8','item9']],[102,['item2','item4']]],columns=['cust_id','Items'])

In [160]: df
Out[160]:
   cust_id                  Items
0      100  [item1, item2, item3]
1      101  [item5, item8, item9]
2      102         [item2, item4]

In [161]: df['Items'] = df['Items'].apply(' '.join)

In [162]: df
Out[162]:
   cust_id                Items
0      100    item1 item2 item3
1      101    item5 item8 item9
2      102          item2 item4

edited Sep 23, 2015 at 13:16

answered Sep 23, 2015 at 12:12

Anand S Kumar

91.4k18 gold badges196 silver badges179 bronze badges

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

7 Comments

EdChum Over a year ago

You can just use df['Items'].apply(' '.join) here`

GeorgeOfTheRF Over a year ago

what does ' ' mean in the apply function?

Anand S Kumar Over a year ago

@mrcet007 its the space, which is being used inbetween the different elements.

GeorgeOfTheRF Over a year ago

Thanks! Can someone also explain why pd.replace doesnt work 7 what is wrong with the code i tried?

Anand S Kumar Over a year ago

because series.replace() is for replacing whole values, what you actually wanted to use was series.str.replace() . but there you cannot use dict, you would have to provide a regular expression or multiple .str.replace() .

|

Collectives™ on Stack Overflow

How to perform string replace operation using pandas in python?

1 Answer 1

7 Comments

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

7 Comments

Related