How does memory allocation work in python in this case?

Question

Suppose I have a very large numpy array a, and I want to add the numerical value 1 to each element of the array. From what I have read so far:

a += 1

is a good way of doing it rather than:

a = a + 1

since in the second case a new array a is created in a different memory slot, while in the first case the old array is effectively replaced in the same memory slot.

Suppose I want to do the following instead:

a = 1-a

What would be the memory efficient way of doing the above?

a is an numpy array? Could you add something like a = numpy.array([1,2,3,4,5,6,7,8]) so we can easily see what you mean? — tdelaney
– tdelaney, Commented May 9, 2018 at 1:21
The documentation for add.at indicates that a += 1 does a certain amount of buffering - docs.scipy.org/doc/numpy-1.13.0/reference/generated/…. Details are buried in numpy code. — hpaulj
– hpaulj, Commented May 9, 2018 at 1:30

user2357112 · Accepted Answer · 2018-05-09 01:10:10Z

6

numpy.subtract(1, a, out=a)

Using the subtract ufunc directly gives you more control than the - operator. Here, we use the out parameter to place the results of the subtraction back into a.

answered May 9, 2018 at 1:10

user2357112

286k32 gold badges490 silver badges571 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

user2357112 Over a year ago

@OlivierMelançon: a is a NumPy array. We're already using NumPy. The cost of importing NumPy is a sunk cost, not an additional cost to pay for using this code.

Julien · Accepted Answer · 2018-05-09 01:04:21Z

3

You could do it in place like so:

a *= -1
a += 1

answered May 9, 2018 at 1:04

Julien

15.3k6 gold badges33 silver badges58 bronze badges

Collectives™ on Stack Overflow

How does memory allocation work in python in this case?

2 Answers 2

1 Comment

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Related