Object identification in Python

Question

In learning Python, I found that when two "names" (or "variables") are assigned to the same value, both of them point to the same memory address. For example

>>> a = 10
>>> b = 10
>>> a is b
True

My question is, when assigning b, how does Python figure out that a 10 object already exists?

One way might be to create the new object for b, scan pre-existing objects in memory to find a duplicate and, if found, point b to it. This sounds expensive, and tricky for more complex objects.

• Daenyth: I found that when two "names" (or "variables") are assigned to the same value, both of them point to the same memory address. -> This is not true.
• Martijn Pieters: See Python "is" operator behaves unexpectedly with integers; you found a CPython implementation detail.

Answer 1

It just keeps a cache of small integer instances.

>>> a = 58435
>>> b = 58435
>>> a is b
False

Precisely how small is easy to determine manually for your implementation. Mine caches −5 to 256.

• wsaleem: That is disappointing - I was hoping for something fancy! By the way, it seems floats are cached too, e.g. a = 10245645646.3; b = 10245645646.3; a is b yields True.
• wsaleem: I think recent numbers are also cached. I am getting caching behavior for numbers well outside my caching range (-5 to 256). e.g. b = 387420489; c = 387420489; b is c is also True. Interestingly, d = 9**9; d is b gives False even though d==b is True.
• Vasiliy Faronov: @wsaleem But if you split into separate lines in the REPL (instead of using the semicolon) you get False. But then if you wrap it in a function, you get True again. Interesting. May have something to do with the interpreter’s optimizations and with local vs. global variables.
• user40980: Related on CodeGolf.SE: Write a program that makes 2 + 2 = 5 the tinkering with the integer cache in particular.

Answer 2

is operator is not the same as ==.

While == checks the value of object, is checks identity of it using it's memory address.

As mentioned before ints from -5 to 256 (inclusive) are special (small) and Python will not recreate those objects. Ints from outside that range will be new objects every time so is will return True only if they are the same exact object.

That happens also to: chars, empty strings and some other strings - that looks like Python identifiers.

(I think that bytes as well, but I'm not sure).

If you want to read more, please take a look the fallowing post: http://mrfuxi.github.io/blog/python-gotchas/