Question

I have a list which contains first and last names, like this one:

names = ["John Smith", "Rob Julian", "Eric Walls"]

I want to get only the first names in this list.

I achieved that by doing:

first_names = [n.split(" ")[0] for n in names]

And that gave me the wanted result.

But in my opinion this is pretty ugly, is there any better way to achieve that goal?

Answer 1

Yes, but not really. Performance wise you are better off with what you have.

first_names = []
for n in names:
    first_names.append(n.split()[0])

would work, but i love list comprehension in python. I mean whats wrong with

first_names = [n.split()[0] for n in names]

For fun, you could also do the following. I would imagine if you are processing a very large list this might have the best performance. But, you might want to investigate that first.

first = lambda n : n.split()[0]
first_names = [first(name) for name in names]

to make this comprehensive, you could also use the lambda to map it.

first = lambda n : n.split()[0]
first_names = map(first,names)

Per comments, I am adding yet another way

from operator import itemgetter
first_names = map(itemgetter(0), map(str.split, names))

in conclusion, yes there are other ways to do it.

but your original seems to be the most preferred. If speed is an issue you may want to tinker with the others.

updated with time

not the most scientific, but using a list of approximaltey 3.5million names, I ran the above calling the files n0-4 and ran time n0;time n1; time n2; time n3; time n4
here are my results. Looks as if the original list comprehension was the fastest on my machine.

real    0m8.433s
user    0m7.064s
sys     0m1.288s

real    0m8.213s
user    0m6.852s
sys     0m1.300s

real    0m8.581s
user    0m7.240s
sys     0m1.264s

real    0m8.374s
user    0m7.164s
sys     0m1.140s

real    0m11.890s
user    0m10.101s
sys     0m1.672s

(I ran them a few times in differing orders, times were consistent.)