What is the purpose of the`reduce` function in Python?

Question

In which cases is useful and how can I use the function reduce?

Answer 1

The function reduce, available in the built-in module functools, serves to "reduce" an iterable (as a list) to a single value.

It is a slightly more common paradigm in functional languages, but it is also quite useful in object-oriented/imperative languages like Python.

But as such, "narrows down" a list?

---

Suppose you have a list minha_lista, that only contains numbers, and that you want the result of multiplication among all these numbers. One of the ways to do this is with a for:

# Queremos 2 * 4 * 5 * 2, que é igual a 80
minha_lista = [2, 4, 5, 2]

produto_total = 1
for numero in minha_lista:
     produto_total *= numero

print(produto_total)  # 80

This works, but can also be done with reduce:

from functools import reduce

minha_lista = [2, 4, 5, 2]

def mult(x, y):
    return x * y

produto_total = reduce(mult, minha_lista)
print(produto_total)  # 80

Or, equivalently but a little more concisely, with a function lambda (which is nothing more than a function defined in a single expression/line):

from functools import reduce

minha_lista = [2, 4, 5, 2]

produto_total = reduce(lambda x, y: x * y, minha_lista)
print(produto_total)  # 80

So what’s going on there? If we take a look at "slow motion," what happens is this:

1. reduce receives as function argument applied to the function mult (or lambda in the second case)
2. reduce receives as eternal the list minha_lista
3. reduce consider how x and y, respectively, the first two elements of minha_lista: 2 and 4
4. reduce applies to x and y the function mult, which multiplies one by the other: in 2 * 4, the result is 8.
5. The result of the application of the function (our 8) becomes x, and the next unprocessed element (5, third element of our list) becomes y
6. The function mult is applied to x and y: 8 * 5 returns 40, and this value returned becomes our new x

7. The function mult is last applied to x and the last item on our list, 2. 40 * 2 returns us 80

8. As there are no more elements to apply the function, reduce returns the final result: 80, as expected.

In short, the function applies another function sequentially and cumulatively to a list (iterator), and "reduces" the list to a single final result.

---

Another educational example is to calculate the minimum value of a list:

from functools import reduce

minha_lista = [2, 4, 5, 2, 1]

def achar_minimo(x, y):
    if x < y:
        return x
    else:
        return y

minimo = reduce(achar_minimo, minha_lista)
print(minimo)  # 1

In this case, the first argument x is used to hold the lowest value seen, and it is compared to other values y of the list.

I say that this example is didactic because there is already an equivalent function in Python: a min. Still, I think the example is nice to give another idea of the power of reduce.

In addition to the above, the reduce also accepts an optional third argument, the initial_value. As the name implies, this argument can be given as initial value, and it is in effect the same thing as putting that value at the beginning of your list.

---

These examples are simple and may seem silly, but as the applied functions become more complex, use the reduce can be a good one to decrease the size of the code and make it clearer. Anyway, it is a good tool to have.