Question

I need function-aggregator, which would reduce two lists to one total number. ‘Items’ is supposed to be vector of Booleans.

So, I wrote these functions:

def element_wise_multiplication(weights, items):
    return map(lambda x, y: x * y, weights, items)

def total(weights, items):
    return sum(element_wise_multiplication(weights, items))

They look OK to me, but the problem is that profiler showed that the line with lambda in it is responsible for 95% of runtime, so its performance is pretty much unacceptable.

What is the most efficient way to implement it?

P.S. I am aware of NumPy’s arrays, but I would like to use PyPy on this one. Or is using it not worth it in this case?

Answer 1

You can take care of that with a generator like so:

from itertools import izip
value = sum((x * y) for x, y in izip(weights, items))

izip accomplishes the same thing as the built-in zip, but without the memory overhead.