Applying sort to DataFrameGroupBy groups in pandas

Question

    brand         vehicleType
0   volkswagen      NaN
1   audi           coupe
2   jeep             suv
3   volkswagen     kleinwagen
4   skoda          kleinwagen
5   bmw             limousine
6   peugeot         cabrio
7   volkswagen      limousine
8   ford            bus
9   volkswagen     kleinwagen
10  mazda          limousine
11  volkswagen     kombi
12  volkswagen     kombi
13  volkswagen     kombi
14  nissan         suv

d=df.groupby(['brand','vehicleType']).vehicleType.count()

From the dataframe, I got the following table by performing the above operation

Now, how do I perform a descending sorting operation to the count column which is unnamed?

AttributeError: 'Series' object has no attribute 'columns', i think this error pops because d is not a dataframe?? — HEMANTH Inukonda Kamalakannan
– HEMANTH Inukonda Kamalakannan, Commented Jul 30, 2017 at 22:36
Yeah. If you could post your original data (as text), that would be great. — cs95
– cs95, Commented Jul 30, 2017 at 22:37
nope it doesn't come ,i just wanted the counted values to be sorted in descending, other than that rest all should be the same — HEMANTH Inukonda Kamalakannan
– HEMANTH Inukonda Kamalakannan, Commented Jul 30, 2017 at 22:54
I assume by "rest all should be the same" you mean you want to preserve the brand order. If that's the case, use d.sort_values(ascending=False).sort_index(level='brand', sort_remaining=False) — user2285236
– user2285236, Commented Jul 30, 2017 at 23:06

cs95 · Accepted Answer · 2017-07-30 23:16:10Z

2

You'll need to call Series.groupby(level=0). To sort each group, you can use a lambda with pd.Series.sort_values:

d.groupby(level=0).apply(lambda x: x.sort_values(ascending=False))

To get the largest item from each group, sorting is no longer necessary. use nlargest(1):

d.groupby(level=0, group_keys=False).apply(pd.Series.nlargest, n=1)

edited Jul 30, 2017 at 23:16

answered Jul 30, 2017 at 23:07

cs95

406k106 gold badges744 silver badges797 bronze badges

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

2 Comments

HEMANTH Inukonda Kamalakannan Over a year ago

What is group_keys? what does it do?

Amit Portnoy Over a year ago

'group_keys=False' tell pandas that you don't want to add the group keys the the resulting dataframe (in this case it makes sense because your dataframe is already grouped with those keys)

Collectives™ on Stack Overflow

Applying sort to DataFrameGroupBy groups in pandas

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related