Factorizing a number in Python

Question

Here\'s my code:

def factorize(n):
    sieve = [True] * (n + 1)

    for x in range(2, int(len(sieve) ** 0.5) + 1):
        if sieve[x]: 
            for i i         


        

Answer 1

The code in the answer by Blender is very nice, but that algorithm is lacking in one very important respect: it tests way too much. E.g. trying to factorize n=392798360393, which is a prime number, it will try to divide it by all the numbers below it (including itself). This will take a lot of time.

Is it really necessary? If n == a*b and a < b, having found a we don't really need to test divide n by b. We know that it too divides n, because n/a == b implies n/b == a. So we only need to test while the potential factor a is smaller than (or equal to) the potential factor b. That is to say, until the square root of the number is reached.

Also, instead of starting over from 2 for each reduced n it's OK to start from the previous value of i:

def factors(n):    # (cf. https://stackoverflow.com/a/15703327/849891)
    j = 2
    while n > 1:
        for i in xrange(j, int(sqrt(n+0.05)) + 1):
            if n % i == 0:
                n /= i ; j = i
                yield i
                break
        else:
            if n > 1:
                yield n; break

Indeed, factoring 9000009 by this code takes 0.08 seconds on Ideone, instead of 0.59 seconds.

This is guaranteed to produce only primes (because we divide out each factor found, and we try the candidates in non-decreasing order). The overhead of generating primes first and then testing only by the primes (a.o.t. testing by all numbers, as done here above) will be worth it if we are factoring several numbers at once; for factoring just one number it might not be worth it, depending on the speed of your prime generation.

---

But then, what really should be done when factoring several numbers at once, is to create the smallest factor sieve first where we mark each number in a given range by its smallest (prime) factor (from which it was generated by the sieve) instead of just by True or False as in the sieve of Eratosthenes. This smallest factor sieve is then used for efficient factorization of each of the numbers given, in the same range, by successive division by their factors, from the smallest up, which are efficiently found by a direct look-up in the sieve instead of testing anew:

def sfs_factorize(nums):
    top = max(nums)
    sv = smallest_factor_sieve(top)
    for k in nums:
        fs = [] ; n = k
        while n > 1:
            f = sv[n]    # no testing
            n /= f
            fs.append(f)
        print( k, list(fs))

Answer 2

My one-liner with python 3.8 (assignment expressions used)

f = lambda n: (p:=[next(i for i in range(2, n+1) if n % i == 0)] if n>1 else [])+(f(n//p[0]) if p else [])

video with details - Factorizing number in Python

Answer 3

The Sieve of Eratosthenes helps you find prime numbers below a certain limit. It's not really going to help you with finding the factors of a particular number.

If you want to do that, the simplest approach that I can see is something like this:

def factors(n):
    while n > 1:
        for i in range(2, n + 1):
            if n % i == 0:
                n /= i
                yield i
                break

for factor in factors(360):
    print factor

This basically finds the smallest factor of n (which is guaranteed to be prime), divides n by that number and repeats the process until n is equal to 1.

The output is:

2
2
2
3
3
5

They multiply out to the original number:

>>> from operator import mul
>>> reduce(mul, factors(360))
360

Answer 4

I'm not familiar enough to tell if this question should be deleted or whatever, but I'll help out anyhow.

You have the sieve part correct, I think. The key is to use a while loop to keep on dividing out a valid prime factor more than once.

factors = []
sieve[0] = sieve[1] = False # So I don't have to worry about trying to skip over these two
for testFactIndex, isPrime in enumerate(sieve):
    if isPrime:
        while n%testFactIndex == 0:
            n = n/testFactIndex
            factors.append(testFactIndex)
return factors