On 10/23/2014 3:39 PM, Ian Kelly wrote:
On Thu, Oct 23, 2014 at 11:07 AM, Seymore4Head
<Seymore4Head@hotmail.invalid> wrote:
BTW I forgot to add that example 2 and 3 don't seem to be too useful
in Python 3, but they are in Python 2. I don't understand how the
Python 3 is an improved version.
In Python 2, range returns a list containing all the requested
elements. This is simple to work with but often not desirable, because
it forces the entire sequence to be loaded into memory at once. Large
ranges may cause excessive memory use and related slowdowns, while
very large ranges may not even be usable.
In Python 3, range returns a compact object that knows what elements
it contains and can iterate over them (which is the most common use of
ranges by far), but without needing to load them all into memory at
once. While iterating, only the current element needs to exist in
memory. You can still get the range as a list if you want it, by
calling list(range(10)), but it doesn't force you to create a list,
which makes it more versatile than the Python 2 construct.
The more specialized data structure used in Python 3 also allows for
certain optimizations, for example membership testing. In Python 2, if
you do the test "97 in range(100)", it has to construct a 100-element
list and then iterate over 97 of the elements before discovering that
the list does in fact contain the number 97. The Python 3 range object
knows what the bounds of the range are, so all it has to do is check
that 0 <= 97 < 100 to know that the number 97 is included in the
range.
Example that works 'instantly' in 3.x, but would be slightly insane in
2.x, even if you have enough memory for it to work on a 64 bit build.
>>> r =range(0, 1000000000000000, 999937)
>>> 3428761974385 in r
False
>>> 3428761974386 in r
True
--
Terry Jan Reedy
--
https://mail.python.org/mailman/listinfo/python-list
