Using 'groupby' without aggregation and sorting within groups

Question

I have countries_of_the_world.csv . Basically, it's the table with the following bits of information:

Country    Region             GDP
Austria    Western Europe     100
Chad       Africa             30

I need to sort GDP values in descending order by region with countries inside these regions. It should look like:

Region     Country        GDP
Africa     Egypt          42
           Chad           30
           Kongo          28
Oceania    Australia      120
           New Zealand    100
           Indonesia      50

I tried 'groupby' but it doesn't work without aggregation function applied so I tried lambda but it didn't sort correctly:

countries.sort_values(['GDP'], ascending=False).groupby(['Region','Country']).aggregate(lambda x:x)

How can I handle it?

Answer 1

Use DataFrame.sort_values by both columns and then convert Region and Country to MultiIndex by DataFrame.set_index :

df1 = (countries.sort_values(['Region','GDP'], ascending=[True, False])
                .set_index(['Region','Country']))
print (df1)
                     GDP
Region  Country         
Africa  Egypt         42
        Chad          30
        Kongo         28
Oceania Australia    120
        New Zealand  100
        Indonesia     50