semantics of [:]

[Luis Zarrabeitia]

On Friday 20 November 2009 11:15:55 am Esmail wrote:
> Could someone help confirm/clarify the semantics of the [:] operator
> in Python?

In short, a[:], when "a" is a list, returns a slice ("sublist") of "a" that 
happens to contain all the elemnts of "a". In general, a[low:upper:step] will 
give you the sublist containing all the items from low to upper, with the 
specified step size. When you omit the three arguments, you get  the whole 
list.

When used on the left side, it means "replace the contents of this slice with 
the contents of this iterable". (The get slice and set slice behaviors are 
defined by the class of «a», so while this explanation applies to list, it 
may not apply, and will likely not apply, to other data structures).

Note that the sublist is always another list (even when you get the full 
slice). So, in your experiments:

> ############# 1 ##################
> b = a[:] # b receives a copy of a, but they are independent

Indeed, a[:] would return a full slice, i.e, another list containing the same 
elements as «a»

> # The following two are equivalent
> ############# 2 ##################
> c = []
> c = a[:] # c receives a copy of a, but they are independent
> ############# 3 ##################
> d = []
> d[:] = a # d receives a copy of a, but they are independent

Not exactly, but the difference is not big in this particular circumstance.

In (2), you create an empty list, discard it immediately, create another list 
a[:] with the contents of «a», and assign it to «c».

In (3) you create an empty list, and then fill it with the elements of «a».

To see the difference, try:

e = []
g = e
e = a[:]
print g

and 

e = []
g = e
e[:] = a
print g

> ### 1 ### is the preferred, shorter way to do what ## 2 ## and ## 3 ##
> do.

Depends on what you want to achieve. Some times, you need to keep the list and 
change the contents. That would be (3). For instance, when updating 
the "dirnames" output of the iterator returned by os.walk . Some times, you 
need to discard de old list, and get a new one. (1) would be the way to go. 
Note that (2) and (1) are the same, in both you discard the old value, if it 
exists, but in (2) you are creating a new list only to discard it right 
away - you shouldn't do that.

> Am I correct with this?

I hope I could help.

-- 
Luis Zarrabeitia (aka Kyrie)
Fac. de Matemática y Computación, UH.
http://profesores.matcom.uh.cu/~kyrie