Alter list items within loop

[Brendan]

Can someone please explain what is happening in the output below? The
number 3 never gets printed. Does Python make a copy of a list before
it iterates through it?: print i
if i == 2 :
e.remove(i)


1
2
4[1, 3, 4]

 

[Emile van Sebille]

On 6/11/2009 11:54 AM Brendan said...

Can someone please explain what is happening in the output below?

you delete e[2] before displaying it.

The

number 3 never gets printed. Does Python make a copy of a list before
it iterates through it?:

No.

Mods to a list while passing it is generally not a good idea. Sometimes
passing the list backwards works.

Emile

print i
if i == 2 :
e.remove(i)


1
2
4[1, 3, 4]

 

[Tim Harig]

Can someone please explain what is happening in the output below? The
number 3 never gets printed. Does Python make a copy of a list before
it iterates through it?:

You can see what is happening by printing the list as you work through the
loop:
.... print e
.... print i
.... if i == 2 :
.... e.remove(i)
....
[1, 2, 3, 4]
1
[1, 2, 3, 4]
2
[1, 3, 4]
4

first loop:
i = 0
e = e[0] = 1

second loop
i = 1
e = e[1] = 2

third loop
i = 2
e = e[2] = 4

number 3 never gets printed. Does Python make a copy of a list before
it iterates through it?:

No, complex types are passed by reference unless explicity copied. You can
do what you want by making an explicit copy before entering the loop:

e = range(1,5)
for i in e[:]:

.... print e
.... print i
.... if i == 2 :
.... e.remove(i)
....
[1, 2, 3, 4]
1
[1, 2, 3, 4]
2
[1, 3, 4]
3
[1, 3, 4]
4

 

[Tim Harig]

*All* types are passed by reference unless explicitly copied. Python does
make special cases for simple and complex types.

That is technically true; but, you will not have this issue with simple
singlular data types. Technically the difference as to whether you will
have this problem depends on whether or not an object is mutable.
Simple objects (numbers and strings) are all immutable. Since this issue
revolves around changing objects in place, it cannot arise with immutable
objects. I am not always conscous of whether I am working with objects
that are mutable or immutable; but, I am generally concious of the general
complexity of the object. Whenever I am working with objects that are
complex, I am reminded to watch out for mutability issues. So, while it is not
totally correct to think of it this way, I find it an easier guideline to
follow.

 

[Lie Ryan]

That is technically true; but, you will not have this issue with simple
singlular data types. Technically the difference as to whether you will
have this problem depends on whether or not an object is mutable.
Simple objects (numbers and strings) are all immutable. Since this issue
revolves around changing objects in place, it cannot arise with immutable
objects. I am not always conscous of whether I am working with objects
that are mutable or immutable; but, I am generally concious of the general
complexity of the object. Whenever I am working with objects that are
complex, I am reminded to watch out for mutability issues. So, while it is not
totally correct to think of it this way, I find it an easier guideline to
follow.

Everything is passed as an object[1], no matter how simple or complex,
mutable or immutable.

This can be seen:
.... print id(a)
....8578336

[1] internally as PyObject pointer but that's implementation detail