0

only mutable objects can be changed in-place; strings, tuples, and other objects will always have to create new objects if you change them. What does in-place mean here,and why tuples, strings are special cases here

phoenixwu0229
  • 11
  • 3
  • in place is _hic et nunc_ - you can manipulate them as you want. del your_list[i] will delete ith item in your list and this will make actual change in your list object. but you cant do the same thing to strings. if you want to change a = "string" to somthing else, you have to create another b variable keeping manipulated string object in it. as you expect a is unchanged. you can just destroy or assign the memory place this variable holds, to the b, like a=b. – marmeladze Apr 13 '15 at 10:57
  • @Martijn: The phrase "in-place" is not explained on the dupe page. – unutbu Apr 13 '15 at 10:58
  • @unutbu: does it really matter here? At issue is understanding what the difference between the two types of objects is. – Martijn Pieters Apr 13 '15 at 11:12
  • @Martijn: The distinction may not matter to you because you know what "in-place" and "mutable" mean, but it may not be so clear to someone trying to learn the meaning of the terms. A person looking for the definition of "in-place" won't find it on the dupe page since the dupe page never mentions the term. Since the terms "in-place" and "mutable" are not fungible, I think it would be nice to address the OP's question directly. – unutbu Apr 13 '15 at 12:54
  • @unutbu: we can do two things here: clean up this question (and tighten it up a bit to focus just on *in-place*, removing the broad *why .. special cases* part), or by improving the dupe to include the term in-place. Personally, I'd prefer the latter, and keep everything in the one place. Update the canonical in-place, as it were.. – Martijn Pieters Apr 13 '15 at 13:56
  • @Martijn: I'm open to either option, but I don't see how to gracefully add the question "what does in-place mean" to the dupe target Q and answer it in the (highest-voted?) A without it sticking out like a wart. – unutbu Apr 13 '15 at 20:19
  • @Martijn: Also, is it fair to add to a question, thus making all answers (except perhaps the edited answer) incomplete? – unutbu Apr 13 '15 at 20:25
  • @unutbu: the top answer cannot be altered a little just to add the term *in place* somewhere? It doesn't have to be added to the question. – Martijn Pieters Apr 13 '15 at 21:28
  • @Martijn: Don't you think an explanation of "in place" should include some discussion of the memory location of the container not changing? Maybe some demonstration using `id`? I don't see how that fits simply into the top-voted answer. Also, I'm not comfortable editing someone else's answer to include a lot of info the original author showed no intention of including himself. – unutbu Apr 13 '15 at 21:51
  • 2
    @unutbu: I see that the top answer *already* uses the phrase "in place". In my mind it just an expression. I'm still not sure that the idiom needs explaining here when the concepts are already well enough covered. Re-evaluating this post and the dupe, I see no point in explaining that separately, really. – Martijn Pieters Apr 13 '15 at 21:54
  • The "in place" should come from the "In-Place Algorithm" in computer science, that is to operate directly on its input and changes it, in here it may mean that to operate on the object without creating a new copy of the object – phchen2 Jul 11 '17 at 00:07

1 Answers1

4

In Python a variable holds a reference to an object.

To "change" a variable, you may either change the referenced object. Or reference an other object.

"Immutable" object force you to use the first solution as the object by itself cannot be modified. "Mutable" objects give you the choice between the two options.

As string are immutable, you cannot "change" a string object. All you can do, is create a new string and update the reference to hold that newly created string:

>>> v1 = v2 = "hello"
>>> v1 += "world"

# v1 and v2 no longer references the same object:
>>> v1 is v2
False

>>> v1
'helloworld'
>>> v2
'hello'

But as list are mutable you have the choice to change them "in place":

>>> v1 = v2 = ['h', 'e', 'l', 'l', 'o']
>>> v1 += ['w', 'o', 'r', 'l', 'd']

# v1 and v2 still references the same object:
>>> v1 is v2
True

>>> v1
['h', 'e', 'l', 'l', 'o', 'w', 'o', 'r', 'l', 'd']
>>> v2
['h', 'e', 'l', 'l', 'o', 'w', 'o', 'r', 'l', 'd']

Or to create a new list and update the reference:

>>> v1 = v2 = ['h', 'e', 'l', 'l', 'o']
>>> v1 = v1 + ['w', 'o', 'r', 'l', 'd']

# v1 and v2 no longer references the same object:
>>> v1 is v2
False

>>> v1
['h', 'e', 'l', 'l', 'o', 'w', 'o', 'r', 'l', 'd']
>>> v2
['h', 'e', 'l', 'l', 'o']
Sylvain Leroux
  • 50,096
  • 7
  • 103
  • 125