I'd like to test whether a term has only one solution.
(Understanding that this might be done in different ways) I've done the following and would like to understand why it doesn't work, if it can be made to work, and if not, what the appropriate implementation would be.
First, I have an "implies" operator (that has seemed to work elsewhere):
:- op(1050,xfy,'==>').
'==>'(A,B) :-·forall(call(A), call(B)).
next I have my singleSolution predicate:
singleSolution(G) :- copy_term(G,G2), (call(G), call(G2)) ==> (G = G2).
Here I'm trying to say: take a term G and make a copy of it, so I can solve them independently. Now if solving both independently implies they are equal, then there must be only one solution.
This works in some simple cases.
BUT.
I have a predicate foo(X,Y,Z) (too large to share) which solves things properly, and for which singleSolution can answer correctly. However, X,Y,Z are not fully ground after singleSolution(foo(X,Y,Z)) is called, even though they would be after directly calling foo(X,Y,Z).
I don't understand that. (As a sanity test: I've verified that I get the same results under swi-prolog and gprolog.)
EDIT: Here is an example of where this fails.
increasing([]).
increasing([_]).
increasing([X,Y|T]) :- X < Y, increasing([Y|T]).
increasingSublist(LL,L) :-·
sublist(L,LL),
length(L, Len),
Len > 1,
increasing(L).
then
| ?- findall(L, singleSolution(increasingSublist([1,2],L)),R).
R = [_]
yes
But we don't know what L is.
This seems to work, but I'm not sure if it's logically sound :)
It uses call_nth/2, a nonstandard but common predicate. It abuses throw to short-circuit the computation. Using bagof/3 instead of findall/3 lets us keep the Goal argument bound (and it will fail where findall/3 would succeed if it finds 0 solutions).
only_once(Goal) :-
catch(bagof(_, only_once_(Goal), _), too_many, fail).
only_once_(Goal) :-
call_nth(Goal, N),
( N > 1
-> throw(too_many)
; true
).
Testing it (on SWI):
?- only_once(member(X, [1])).
X = 1.
?- only_once(member(a, [a, b])).
true.
?- only_once(member(X, [a, b])).
false.
?- only_once(between(1,inf,X)).
false.
Unfortunately, I don't think call_nth/2 is supported in GNU Prolog.
Another possible solution:
single_solution(G) :-
copy_term(G, H),
call(G),
!,
( ground(H)
-> true
; \+ ( call(H), G \= H ) % There is no H different from G
).
p(a).
p(a).
q(b).
q(c).
Examples:
?- single_solution( p(X) ).
X = a.
?- single_solution( q(X) ).
false.
?- single_solution( member(X, [a,a,a]) ).
X = a.
?- single_solution( member(X, [a,b,c]) ).
false.
?- single_solution( repeat ).
true.
?- single_solution( between(1,inf,X) ).
false.
?- single_solution( between(1,inf,5) ).
true.
Here is an another approach I came up with after #gusbro commented that forall/2 doesn't bind variables from the calling goal.
single_solution(G) :-·
% duplicate the goal so we can solve independently
copy_term(G,G2),
% solve the first goal at least / at most once.
G, !,
% can we solve the duplicate differently?
% if so, cut & fail. Otherwise, succeed.
(G2, G2 \= G, !, fail; true).
This one tickled my interest in theory:
Is it possible to write an inconsistent Prolog program, i.e. a program that answers both false and true depending on how it is queried, using only pure Prolog, the cut, and false?
For example, one could query p(1) and the Prolog Processor would says false. But when one queries p(X) the Prolog Processor would give the set of answers 1, 2, 3.
This can be easily achieved with "computational state examination predicates" like var/1 (really better called fresh/1) + el cut:
p(X) :- nonvar(X),!,member(X,[2,3]).
p(X) :- member(X,[1,2,3]).
Then
?- p(1).
false.
?- p(X).
X = 1 ;
X = 2 ;
X = 3.
"Ouch time" ensues if this is high-assurance software. Naturally, any imperative program has no problem going off the rails like this on every other line.
So. can be done without those "computational state examination predicates"?
P.S.
The above illustrates that all the predicates of Prolog are really carrying a threaded hidden argument of the "computational state":
p(X,StateIn,StateOut).
which can be used to explain the behavour of var/1 and friends. The Prolog program is then "pure" when it only calls predicates that neither consult not modify that State. Well, at least that seems to be a good way to look at what is going on. I think.
Here's a very simple one:
f(X,X) :- !, false.
f(0,1).
Then:
| ?- f(0,1).
yes
| ?- f(X,1).
no
| ?- f(0,Y).
no
So Prolog claims there are no solutions to the queries with variables, although f(0,1) is true and would be a solution to both.
Here is one attempt. The basic idea is that X is a variable iff it can be unified with both a and b. But of course we can't write this as X = a, X = b. So we need a "unifiable" test that succeeds without binding variables like =/2 does.
First, we need to define negation ourselves, since it's impure:
my_not(Goal) :-
call(Goal),
!,
false.
my_not(_Goal).
This is only acceptable if your notion of pure Prolog includes call/1. Let's say that it does :-)
Now we can check for unifiability by using =/2 and the "not not" pattern to preserve success while undoing bindings:
unifiable(X, Y) :-
my_not(my_not(X = Y)).
Now we have the tools to define var/nonvar checks:
my_var(X) :-
unifiable(X, a),
unifiable(X, b).
my_nonvar(X) :-
not(my_var(X)).
Let's check this:
?- my_var(X).
true.
?- my_var(1).
false.
?- my_var(a).
false.
?- my_var(f(X)).
false.
?- my_nonvar(X).
false.
?- my_nonvar(1).
true.
?- my_nonvar(a).
true.
?- my_nonvar(f(X)).
true.
The rest is just your definition:
p(X) :-
my_nonvar(X),
!,
member(X, [2, 3]).
p(X) :-
member(X, [1, 2, 3]).
Which gives:
?- p(X).
X = 1 ;
X = 2 ;
X = 3.
?- p(1).
false.
Edit: The use of call/1 is not essential, and it's interesting to write out the solution without it:
not_unifiable(X, Y) :-
X = Y,
!,
false.
not_unifiable(_X, _Y).
unifiable(X, Y) :-
not_unifiable(X, Y),
!,
false.
unifiable(_X, _Y).
Look at those second clauses of each of these predicates. They are the same! Reading these clauses declaratively, any two terms are not unifiable, but also any two terms are unifiable! Of course you cannot read these clauses declaratively because of the cut. But I find this especially striking as an illustration of how catastrophically impure the cut is.
Given a set of facts that represent parent-child relationships via the predicate parent/2, what are the differences when defining the relation "progenitor"(ancestor) with the predicates pred1/2 and pred2/2 as shown below?
pred1(X,Z) :- parent(X,Z).
pred1(X,Z) :- parent(X,Y), pred1(Y,Z).
pred2(X,Z) :- parent(X,Z).
pred2(X,Z) :- parent(Y,Z), pred2(X,Y).
The major difference are the termination properties of both, provided there are some cycles in the relation. To understand this, I will use a failure-slice which cuts down the program to its essence w.r.t. termination.
pred1(X,Z) :- false, parent(X,Z).
pred1(X,Z) :- parent(X,Y), pred1(Y,Z). % Z has no influence
pred2(X,Z) :- false, parent(X,Z).
pred2(X,Z) :- parent(Y,Z), pred2(X,Y). % X has no influence
For pred1/2, the second argument has no influence on termination at all, whereas in pred2/2, the first argument has no influence. To see this, try the original definitions with the fact:
parent(a,a).
?- pred1(b, _), false.
false. % terminates
?- pred2(b, _), false.
loops.
?- pred1(_, b), false.
loops.
?- pred2(_, b), false.
false. % terminates
See closure/3 for a general approach to avoid such loops.
For completeness, here is another variation of transitive closure which has some conceptual advantages:
pred3(X,Z) :- parent(X,Z).
pred3(X,Z) :- pred3(X,Y), pred3(Y,Z).
Alas, its termination properties are worst. In fact, it never terminates, as is testified by the following fragment:
pred3(X,Z) :- false, parent(X,Z).
pred3(X,Z) :- pred3(X,Y), false, pred3(Y,Z).
In this fragment, the first argument is only passed on. So, no matter, what the arguments are, the program will always loop!
?- pred3(b,b), false.
loops.
I'm trying to figure out a way to check if two lists are equal regardless of their order of elements.
My first attempt was:
areq([],[]).
areq([],[_|_]).
areq([H1|T1], L):- member(H1, L), areq(T1, L).
However, this only checks if all elements of the list on the left exist in the list on the right; meaning areq([1,2,3],[1,2,3,4]) => true. At this point, I need to find a way to be able to test thing in a bi-directional sense. My second attempt was the following:
areq([],[]).
areq([],[_|_]).
areq([H1|T1], L):- member(H1, L), areq(T1, L), append([H1], T1, U), areq(U, L).
Where I would try to rebuild the lest on the left and swap lists in the end; but this failed miserably.
My sense of recursion is extremely poor and simply don't know how to improve it, especially with Prolog. Any hints or suggestions would be appreciated at this point.
As a starting point, let's take the second implementation of equal_elements/2 by #CapelliC:
equal_elements([], []).
equal_elements([X|Xs], Ys) :-
select(X, Ys, Zs),
equal_elements(Xs, Zs).
Above implementation leaves useless choicepoints for queries like this one:
?- equal_elements([1,2,3],[3,2,1]).
true ; % succeeds, but leaves choicepoint
false.
What could we do? We could fix the efficiency issue by using
selectchk/3 instead of
select/3, but by doing so we would lose logical-purity! Can we do better?
We can!
Introducing selectd/3, a logically pure predicate that combines the determinism of selectchk/3 and the purity of select/3. selectd/3 is based on
if_/3 and (=)/3:
selectd(E,[A|As],Bs1) :-
if_(A = E, As = Bs1,
(Bs1 = [A|Bs], selectd(E,As,Bs))).
selectd/3 can be used a drop-in replacement for select/3, so putting it to use is easy!
equal_elementsB([], []).
equal_elementsB([X|Xs], Ys) :-
selectd(X, Ys, Zs),
equal_elementsB(Xs, Zs).
Let's see it in action!
?- equal_elementsB([1,2,3],[3,2,1]).
true. % succeeds deterministically
?- equal_elementsB([1,2,3],[A,B,C]), C=3,B=2,A=1.
A = 1, B = 2, C = 3 ; % still logically pure
false.
Edit 2015-05-14
The OP wasn't specific if the predicate
should enforce that items occur on both sides with
the same multiplicities.
equal_elementsB/2 does it like that, as shown by these two queries:
?- equal_elementsB([1,2,3,2,3],[3,3,2,1,2]).
true.
?- equal_elementsB([1,2,3,2,3],[3,3,2,1,2,3]).
false.
If we wanted the second query to succeed, we could relax the definition in a logically pure way by using meta-predicate
tfilter/3 and
reified inequality dif/3:
equal_elementsC([],[]).
equal_elementsC([X|Xs],Ys2) :-
selectd(X,Ys2,Ys1),
tfilter(dif(X),Ys1,Ys0),
tfilter(dif(X),Xs ,Xs0),
equal_elementsC(Xs0,Ys0).
Let's run two queries like the ones above, this time using equal_elementsC/2:
?- equal_elementsC([1,2,3,2,3],[3,3,2,1,2]).
true.
?- equal_elementsC([1,2,3,2,3],[3,3,2,1,2,3]).
true.
Edit 2015-05-17
As it is, equal_elementsB/2 does not universally terminate in cases like the following:
?- equal_elementsB([],Xs), false. % terminates universally
false.
?- equal_elementsB([_],Xs), false. % gives a single answer, but ...
%%% wait forever % ... does not terminate universally
If we flip the first and second argument, however, we get termination!
?- equal_elementsB(Xs,[]), false. % terminates universally
false.
?- equal_elementsB(Xs,[_]), false. % terminates universally
false.
Inspired by an answer given by #AmiTavory, we can improve the implementation of equal_elementsB/2 by "sharpening" the solution set like so:
equal_elementsBB(Xs,Ys) :-
same_length(Xs,Ys),
equal_elementsB(Xs,Ys).
To check if non-termination is gone, we put queries using both predicates head to head:
?- equal_elementsB([_],Xs), false.
%%% wait forever % does not terminate universally
?- equal_elementsBB([_],Xs), false.
false. % terminates universally
Note that the same "trick" does not work with equal_elementsC/2,
because of the size of solution set is infinite (for all but the most trivial instances of interest).
A simple solution using the sort/2 ISO standard built-in predicate, assuming that neither list contains duplicated elements:
equal_elements(List1, List2) :-
sort(List1, Sorted1),
sort(List2, Sorted2),
Sorted1 == Sorted2.
Some sample queries:
| ?- equal_elements([1,2,3],[1,2,3,4]).
no
| ?- equal_elements([1,2,3],[3,1,2]).
yes
| ?- equal_elements([a(X),a(Y),a(Z)],[a(1),a(2),a(3)]).
no
| ?- equal_elements([a(X),a(Y),a(Z)],[a(Z),a(X),a(Y)]).
yes
In Prolog you often can do exactly what you say
areq([],_).
areq([H1|T1], L):- member(H1, L), areq(T1, L).
bi_areq(L1, L2) :- areq(L1, L2), areq(L2, L1).
Rename if necessary.
a compact form:
member_(Ys, X) :- member(X, Ys).
equal_elements(Xs, Xs) :- maplist(member_(Ys), Xs).
but, using member/2 seems inefficient, and leave space to ambiguity about duplicates (on both sides). Instead, I would use select/3
?- [user].
equal_elements([], []).
equal_elements([X|Xs], Ys) :-
select(X, Ys, Zs),
equal_elements(Xs, Zs).
^D here
1 ?- equal_elements(X, [1,2,3]).
X = [1, 2, 3] ;
X = [1, 3, 2] ;
X = [2, 1, 3] ;
X = [2, 3, 1] ;
X = [3, 1, 2] ;
X = [3, 2, 1] ;
false.
2 ?- equal_elements([1,2,3,3], [1,2,3]).
false.
or, better,
equal_elements(Xs, Ys) :- permutation(Xs, Ys).
The other answers are all elegant (way above my own Prolog level), but it struck me that the question stated
efficient for the regular uses.
The accepted answer is O(max(|A| log(|A|), |B|log(|B|)), irrespective of whether the lists are equal (up to permutation) or not.
At the very least, it would pay to check the lengths before bothering to sort, which would decrease the runtime to something linear in the lengths of the lists in the case where they are not of equal length.
Expanding this, it is not difficult to modify the solution so that its runtime is effectively linear in the general case where the lists are not equal (up to permutation), using random digests.
Suppose we define
digest(L, D) :- digest(L, 1, D).
digest([], D, D) :- !.
digest([H|T], Acc, D) :-
term_hash(H, TH),
NewAcc is mod(Acc * TH, 1610612741),
digest(T, NewAcc, D).
This is the Prolog version of the mathematical function Prod_i h(a_i) | p, where h is the hash, and p is a prime. It effectively maps each list to a random (in the hashing sense) value in the range 0, ...., p - 1 (in the above, p is the large prime 1610612741).
We can now check if two lists have the same digest:
same_digests(A, B) :-
digest(A, DA),
digest(B, DB),
DA =:= DB.
If two lists have different digests, they cannot be equal. If two lists have the same digest, then there is a tiny chance that they are unequal, but this still needs to be checked. For this case I shamelessly stole Paulo Moura's excellent answer.
The final code is this:
equal_elements(A, B) :-
same_digests(A, B),
sort(A, SortedA),
sort(B, SortedB),
SortedA == SortedB.
same_digests(A, B) :-
digest(A, DA),
digest(B, DB),
DA =:= DB.
digest(L, D) :- digest(L, 1, D).
digest([], D, D) :- !.
digest([H|T], Acc, D) :-
term_hash(H, TH),
NewAcc is mod(Acc * TH, 1610612741),
digest(T, NewAcc, D).
One possibility, inspired on qsort:
split(_,[],[],[],[]) :- !.
split(X,[H|Q],S,E,G) :-
compare(R,X,H),
split(R,X,[H|Q],S,E,G).
split(<,X,[H|Q],[H|S],E,G) :-
split(X,Q,S,E,G).
split(=,X,[X|Q],S,[X|E],G) :-
split(X,Q,S,E,G).
split(>,X,[H|Q],S,E,[H|G]) :-
split(X,Q,S,E,G).
cmp([],[]).
cmp([H|Q],L2) :-
split(H,Q,S1,E1,G1),
split(H,L2,S2,[H|E1],G2),
cmp(S1,S2),
cmp(G1,G2).
A simple solution using cut.
areq(A,A):-!.
areq([A|B],[C|D]):-areq(A,C,D,E),areq(B,E).
areq(A,A,B,B):-!.
areq(A,B,[C|D],[B|E]):-areq(A,C,D,E).
Some sample queries:
?- areq([],[]).
true.
?- areq([1],[]).
false.
?- areq([],[1]).
false.
?- areq([1,2,3],[3,2,1]).
true.
?- areq([1,1,2,2],[2,1,2,1]).
true.