CODEWITHSUNDEEP
pythonpython-2.7pandas
Marc Zaharescu
Marc Zaharescu

Reputation: 639

Python: transform data set

I have the following data set

            id  type  value
0             1   A    10
1             1   C   120
2             2   B    20
3             2   C    40
4             3   A    10
5             3   B    50

I want in python to transform it to be like (1,A,10,C,120) (2,B,20,C,40) (3,A,10,B,50)

Any suggestion would be much appreciated

Upvotes: 2

Views: 864

Answers (2)

jezrael
jezrael

Reputation: 862801

You can use:

L = df.groupby('id').apply(lambda x: tuple([x['id'].iat[0]] + 
                                          x[['type','value']].values.flatten().tolist()))
                    .tolist()
print (L)
[(1, 'A', 10, 'C', 120), (2, 'B', 20, 'C', 40), (3, 'A', 10, 'B', 50)]

Upvotes: 2

Nickil Maveli
Nickil Maveli

Reputation: 29711

Perform groupby w.r.t id column. Iterate over each group by converting the other two columns to a list and add the unique value corresponding to the id per group number alongside. Finally, convert them into a tuple and append these to a list.

grouped = df.groupby('id')
L = []
for _, grp in grouped:
    L.append(tuple(grouped.get_group(_)['id'].unique().tolist() + grp[['type','value']].values.ravel().tolist()))
print(L)
#[(1, 'A', 10, 'C', 120), (2, 'B', 20, 'C', 40), (3, 'A', 10, 'B', 50)]

Upvotes: 2

Related Questions