How do you join multiple rows into one row in pandas?-CodePudding

I have a list that I'm trying to add to a dataframe. It looks something like this:

list_one = ['apple','banana','cherry',' ', 'grape', 'orange', 'pineapple','']

If I add the list to a dataframe, using df = pd.DataFrame({'list_one':list_one}) it'll look like this:

       list_one
   -------------
   0   apple
   1   banana 
   2   cherry
   3  
   4   grape
   5   orange
   6   pineapple
   7

I want the combine some of the rows into one row, so that the dataframe looks something like this:

       list_one
   -----------------------------
   0   apple, banana, cherry 
   1   grape, orange, pineapple

Is there a simple way to do this?

Thank you for taking the time to read my question and help in any way you can.

CodePudding user response：

Create mask for match words by Series.str.contains, invert by ~ and crate groups by Series.cumsum, filter only matched rows and pass to GroupBy.agg with join function:

m = df['list_one'].str.contains('\w ')
df = df[m].groupby((~m).cumsum(), as_index=False).agg(', '.join)
print (df)
                   list_one
0     apple, banana, cherry
1  grape, orange, pineapple

CodePudding user response：

Try with groupby and agg:

>>> df.groupby(df.loc[df['list_one'].str.contains('\w ')].index.to_series().diff().ne(1).cumsum(), as_index=False).agg(', '.join)
                   list_one
0     apple, banana, cherry
1  grape, orange, pineapple
>>>