CODEWITHSUNDEEP
pythonpandasaggregaterank
tadalendas
tadalendas

Reputation: 1541

Pandas groupby then rank: correct syntax to specify order when passing dict of parameters

I'd like to specify rank order in pandas. I know you can do it when you do:

df.groupby([x, y]).rank(ascending=False)

Question 1

How can I specify the ranking order (ascending=False) inside agg method:

df.groupby([x, y]).agg({
        ('r', 'c'): 'rank'
    })

Question 2

df.groupby([x, y])['r', 'c'].rank(ascending=False)

Throws an error: KeyError: Columns not found

How can I specify a multi lelvel columns for rank?

Data Structure is as follows:

                                            r 
                                            c 
z         x                           y                             
1         2016-11-01 00:00:00+00:00   3121  143       
                                      3923  11      
                                      3953  4    
                                      4880  12

Upvotes: 2

Views: 344

Answers (1)

jezrael
jezrael

Reputation: 862691

I think you can use:

x = 'x'
y = 'y'
b = df[('r','c')].groupby(level=[x, y]).rank(ascending=False)
print (b)
z  x                          y   
1  2016-11-01 00:00:00+00:00  3121    1.0
                              3923    1.0
                              3953    1.0
                              4880    1.0
Name: (r, c), dtype: float64

Or need tuple - add , to the end:

x = 'x'
y = 'y'
b = df.groupby(level=[x, y])[('r','c'), ].rank(ascending=False)
print (b)
                                    r
                                    c
z x                         y        
1 2016-11-01 00:00:00+00:00 3121  1.0
                            3923  1.0
                            3953  1.0
                            4880  1.0
#print (df)

Upvotes: 1

Related Questions