I'm trying to find the last item (all numbers) without "ruining" the list.
What I currently have is this:
max([_|Xs]-Xs,R):-
Xs \= [],
max2(Xs,R),!.
max([X|Xs]-Xs,R):-
R = X.
max([]-[],99999).
Max2 is a function that finds the last item in an ordinary list:
max2([X],X):-
number(X),!.
max2([_|Xs],R):-
max2(Xs,R),!.
When I try on a list with one item, it works - otherwise fails:
max([22|X]-X,R)
R = 22
max([22,27|X]-X,R)
Stack limit (0.2Gb) exceeded
Stack sizes: local: 0.2Gb, global: 21Kb, trail: 3Kb
Stack depth: 1,560,328, last-call: 0%, Choice points: 1,560,312
Probable infinite recursion (cycle):
[1,560,328] max2([cyclic list], _1452)
[1,560,327] max2([cyclic list], _1484)
I tried other ways, but then when I gave it a list with one item, it converted it:
max([22|X]-X,R)
X = []
R = 22
So I couldn't keep using the X as a free variable.
I hope I wrote it clearly. Many thanks in advance.
In the Logtalk library, I use the following definition:
last(List-Back, Last) :-
List \== Back,
List = [Head| Tail],
last(Tail-Back, Head, Last).
last(List, Last, Last) :-
unify_with_occurs_check(List, Back-Back).
last(List-Back, _, Last) :-
List \== Back,
List = [Head| Tail],
last(Tail-Back, Head, Last).
Sample calls:
?- last([22|X]-X,R).
R = 22 ;
false.
?- last([a,b,c|X]-X,R).
R = c ;
false.
?- last(DL, Last).
DL = [Last|_1648]-_1648 ;
DL = [_1652, Last|_1648]-_1648 ;
DL = [_1652, _1664, Last|_1648]-_1648 ;
DL = [_1652, _1664, _1676, Last|_1648]-_1648 ;
DL = [_1652, _1664, _1676, _1688, Last|_1648]-_1648
...
This definition is quite general as it also works with lists of variables. For example:
?- last([X,Y,Z|Tail]-Tail, Last).
Z = Last ;
false.
In your case, if you can ensure that the difference lists are always well formed and their elements are ground, you can simplify the definition.