Question

Given a list of dictionaries like this:

x = [
        {'name':'a', 'student': 1 , 'age':19}, 
        {'name':'b', 'student': 0 , 'age':10}
    ]

I want to sort it by age only if student is equal to 1. Can I somehow put that if in the following statement?

sortedlist = sorted(x, key=lambda k: k['age'])

Thanks,

Answer 1

If you use itemgetter + a generator, instead of a lambda + list comp, you get the best performance I have found so far. This was tested on a dicts list of 10k elements. Almost a 30% speed increase over list comp + lambda. Also, if you can safely assume ‘student’ is always a valid key and access it directly, you again gain more speed over having to use d.get('student', 0) == 1

from operator import itemgetter

sorted((d for d in x if d['student']==1), key=itemgetter('age'))

• Note about lambda vs itemgetter: The reason itemgetter is faster (and I am mostly sure about this) is because the lookup is done on the C side of code. Whereas when you use a lambda you are doing it on the python side which is slower.