So I've been playing around with python and noticed something that seems a bit odd. The semantics of -1 in selecting from a list don't seem to be consistent.
So I have a list of numbers
ls = range(1000)
The last element of the list if of course ls[-1] but if I take a sublist of that so that I get everything from say the midpoint to the end I would do
ls[500:-1]
but this does not give me a list containing the last element in the list, but instead a list containing everything UP TO the last element. However if I do
ls[0:10]
I get a list containing also the tenth element (so the selector ought to be inclusive), why then does it not work for -1.
I can of course do ls[500:] or ls[500:len(ls)] (which would be silly). I was just wondering what the deal with -1 was, I realise that I don't need it there.
6 Answers
In list[first:last], last is not included.
The 10th element is ls[9], in ls[0:10] there isn't ls[10].
If you want to get a sub list including the last element, you leave blank after colon:
>>> ll=range(10)
>>> ll
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
>>> ll[5:]
[5, 6, 7, 8, 9]
>>> ll[:]
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
I get consistent behaviour for both instances:
>>> ls[0:10]
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
>>> ls[10:-1]
[10, 11, 12, 13, 14, 15, 16, 17, 18]
Note, though, that tenth element of the list is at index 9, since the list is 0-indexed. That might be where your hang-up is.
In other words, [0:10] doesn't go from index 0-10, it effectively goes from 0 to the tenth element (which gets you indexes 0-9, since the 10 is not inclusive at the end of the slice).
It seems pretty consistent to me; positive indices are also non-inclusive. I think you're doing it wrong. Remembering that range() is also non-inclusive, and that Python arrays are 0-indexed, here's a sample python session to illustrate:
>>> d = range(10)
>>> d
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
>>> d[9]
9
>>> d[-1]
9
>>> d[0:9]
[0, 1, 2, 3, 4, 5, 6, 7, 8]
>>> d[0:-1]
[0, 1, 2, 3, 4, 5, 6, 7, 8]
>>> len(d)
10
when slicing an array;
ls[y:x]
takes the slice from element y upto and but not including x. when you use the negative indexing it is equivalent to using
ls[y:-1] == ls[y:len(ls)-1]
so it so the slice would be upto the last element, but it wouldn't include it (as per the slice)
-1 isn't special in the sense that the sequence is read backwards, it rather wraps around the ends. Such that minus one means zero minus one, exclusive (and, for a positive step value, the sequence is read "from left to right".
so for i = [1, 2, 3, 4], i[2:-1] means from item two to the beginning minus one (or, 'around to the end'), which results in [3].
The -1th element, or element 0 backwards 1 is the last 4, but since it's exclusive, we get 3.
I hope this is somewhat understandable.
