[ 常見問題 ] Concatenate strings from several rows using Pandas groupby

Source From Here
Question
I want to merge several strings in a dataframe based on a groupedby in Pandas.

This is my code so far:

view plaincopy to clipboardprint?

1. from io import StringIO  
2.   
3. data = StringIO("""  
4. "name1","hej","2014-11-01"  
5. "name1","du","2014-11-02"  
6. "name1","aj","2014-12-01"  
7. "name1","oj","2014-12-02"  
8. "name2","fin","2014-11-01"  
9. "name2","katt","2014-11-02"  
10. "name2","mycket","2014-12-01"  
11. "name2","lite","2014-12-01"  
12. """)  
13.   
14. # load string as stream into dataframe  
15. df = pd.read_csv(data,header=0, names=["name","text","date"],parse_dates=[2])  
16.   
17. # add column with month  
18. df["month"] = df["date"].apply(lambda x: x.month)  

I don't get how I can use groupby and apply some sort of concatenation of the strings in the column "text". Any help appreciated!

How-To
You can groupby the 'name' and 'month' columns, then call transform which will return data aligned to the original df and apply a lambda where we join the text entries:

view plaincopy to clipboardprint?

1. df['text'] = df[['name','text','month']].groupby(['name','month'])['text'].transform(lambda x: ','.join(x))  

view plaincopy to clipboardprint?

1. df = df[['name','text','month']].drop_duplicates()  

Actually I can just call apply and then reset_index:

view plaincopy to clipboardprint?

1. df.groupby(['name','month'])['text'].apply(lambda x: ','.join(x)).reset_index()  

This message was edited 7 times. Last update was at 11/03/2020 19:28:44