CODEWITHSUNDEEP
pythonpandasdataframe
rosefun
rosefun

Reputation: 1857

How to select a column element according to the size of another column elements?

I have a DataFrame named a. I want to get the top most used time apps.

import pandas as pd 
a=pd.DataFrame({'user':[1,1,1,2,2,2,2],'app':['k','p','s','k','p','s','t'],'time':[5,10,15,10,5,3,1]})

Input:

      user   app   time
0        1      k     5
1        1      p    10
2        1      s    15
3        2      k    10
4        2      p     5
5        2      s     3
6        2      t     1

For example, I want to get the top two most used apps according the column time. I expect the output as follows.

Expected:

      user top1_app top2_app
0        1      s     p
1        2      k     p

As you see, user 1 has the longest time to use the app called s, and has the second longest time to use the app called p.

Hopefully for help and thanks!

Upvotes: 2

Views: 59

Answers (1)

Vaishali
Vaishali

Reputation: 38415

You can rank the time column and then reshape

a['time1'] = a.groupby('user').time.rank(method = 'dense', ascending = False).map({1.0 : 'top1_app', 2.0 : 'top2_app'})

a = a.dropna().pivot('user', 'time1', 'app')
a.columns.name = None
a.reset_index(inplace = True)


    user    top1_app    top2_app
0   1       s           p
1   2       k           p

Upvotes: 3

Related Questions