Pandas: reverse rows with same values

Question

I have a DataFrame, where rows are 'grouped' by the third column (rows in one 'group' have the same value at the third column):

But the values in the second column have a wrong order. I need to reverse rows in each 'group', so DataFrame should look like this:

Is there an effective way to transform the first DataFrame to the second one with pandas?

UPD: Updated with more clear example. The values should be exactly reversed, not just became located in the alphabetical order.

jezrael · Accepted Answer · 2017-03-31 07:42:20Z

3

It seems you need sort_values:

df = df.sort_values(['c3','c2'])
print (df)
   c1   c2  c3
1   1    a   1
0   0    b   1
4   4   aa   2
3   3   bb   2
2   2   cc   2
6   6  xxx   3
5   5  zzz   3

EDIT:

You can use groupby and change order by [::-1]:

#more general solution, working with not unique index also
def reversing(x):
    x['c2'] = x['c2'].iloc[::-1].values
    return x

df = df.groupby('c3', sort=False)).apply(reversing)
print (df)
   c1 c2  c3
0   0  r   1
1   1  b   1
2   2  n   2
3   3  x   2
4   4  f   2
5   5  f   3
6   6  r   3

#solution working only with unique monotonic increasing index, because reset index
df['c2'] = df.groupby('c3', sort=False)['c2'].apply(lambda x: x[::-1]).reset_index(drop=True)
print (df)
   c1 c2  c3
0   0  r   1
1   1  b   1
2   2  n   2
3   3  x   2
4   4  f   2
5   5  f   3
6   6  r   3

Solutions, where order of values in c1 is changed:

You can sorting by index (has to be unique monotonic increasing).

df=df.reset_index().sort_values(['c3','index'],ascending=[True, False]).drop('index',axis=1)
print (df)
   c1 c2  c3
1   1  r   1
0   0  b   1
4   4  n   2
3   3  x   2
2   2  f   2
6   6  f   3
5   5  r   3

If column c1 is unique monotonic increasing (0,1,2..) as default index:

df = df.sort_values(['c3','c1'], ascending=[True, False])
print (df)
   c1 c2  c3
1   1  r   1
0   0  b   1
4   4  n   2
3   3  x   2
2   2  f   2
6   6  f   3
5   5  r   3

edited Mar 31, 2017 at 7:42

answered Mar 30, 2017 at 11:48

jezrael

868k103 gold badges1.4k silver badges1.3k bronze badges

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

1 Comment

Amir Over a year ago

Thanks for the answer, but these letters are just an example of values, sorry that I've incorrectly cited this example. In my case the values in the groups shouldn't be in alphabetical order, they should be exactly reversed. I've updated the post with more clear example.

Collectives™ on Stack Overflow

Pandas: reverse rows with same values

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related