Why does my Prolog predicate invert/2 not work? - list

I'm new to Prolog and as an exercise I want to make an list invertion predicate. It uses the add_tail predicate that I made earlier—some parts might be redundant, but I don't care:
add_tail(A, [], A) :-
!.
add_tail([A|[]], H, [A,H]) :-
!.
add_tail([A|B], H, [A|C]) :-
add_tail(B,H,C).
It works same as builtin predicate append/3:
?- add_tail([a,b,c], d, A).
A = [a, b, c, d].
?- append([a,b,c], [d], A).
A = [a, b, c, d].
When I use append in my invert predicate, it works fine, but if I use add_tail, it fails:
invert([], []).
invert([A|B], C) :-
invert(B, D),
append(D, [A], C).
invert2([], []).
invert2([A|B], C) :-
invert2(B, D),
add_tail(D, A, C).
?- invert([a,b,c,d], A).
A = [d, c, b, a].
?- invert2([a,b,c,d], A).
false. % expected answer A = [d,c,b,a], like above
What exactly is my mistake? Thank you!

The implementation of add_tail/3 does not quite behave the way you expect it to.
Consider:
?- append([], [d], Xs).
Xs = [d].
?- add_tail([], d, Xs).
false.
That's bad... But it gets worse! There are even more issues with the code you presented:
By using (!)/0 you needlessly limit the versatility of your predicate.
Even though [A|[]] maybe correct, it obfuscates your code. Use [A] instead!
add_tail is a bad name for a predicate that works in more than one direction.
The variable names could be better, too! Why not use more descriptive names like As?
Look again at the variables you used in the last clause of add_tail/3!
add_tail([A|B], H, [A|C]) :-
add_tail(B, H, C).
Consider the improved variable names:
add_tail([A|As], E, [A|Xs]) :-
add_tail(As, E, Xs).
I suggest starting over like so:
list_item_appended([], X, [X]).
list_item_appended([E|Es], X, [E|Xs]) :-
list_item_appended(Es, X, Xs).
Let's put list_item_appended/3 to use in list_reverted/2!
list_reverted([], []).
list_reverted([E|Es], Xs) :-
list_reverted(Es, Fs),
list_item_appended(Fs, E, Xs).
Sample query:
?- list_reverted([a,b,c,d], Xs).
Xs = [d, c, b, a].

It is difficult to pinpoint your exact mistake, but the first two clauses of add_tail/3, the ones with the cuts, are wrong (unless I am misunderstanding what the predicate is supposed to do). Already the name is a bit misleading, and you should should care that you have redundant code.
list_back([], B, [B]).
list_back([X|Xs], B, [X|Ys]) :-
list_back(Xs, B, Ys).
This is a drop-in replacement for your add_tail/3 in your definition of invert/2. But as you are probably aware, this is not a very clever way of reversing a list. The textbook example of how to do it:
list_rev(L, R) :-
list_rev_1(L, [], R).
list_rev_1([], R, R).
list_rev_1([X|Xs], R0, R) :-
list_rev_1(Xs, [X|R0], R).

First try the most general query, to see which solutions exist in the most general case:
?- add_tail(X, Y, Z).
yielding the single answer:
X = Z,
Y = []
That's probably not the relation you intend to define here.
Hint: !/0 typically destroys all logical properties of your code, including the ability to use your predicates in all directions.

The first clause of add_tail/3 has a list as second argument, so it will never apply to your test case. Then we are left with 2 clauses (simplified)
add_tail([A],H,[A,H]):-!.
add_tail([A|B],H,[A|C]) :- add_tail(B,H,C).
You can see that we miss a matching clause for the empty list as first argument. Of course, append/3 instead has such match.

based on previous answer of "#mat" the problem is residue in the first two lines
your predicate add_tail is not like append because
with append i get this
| ?- append(X,Y,Z).
Z = Y,
X = [] ? ;
X = [_A],
Z = [_A|Y] ? ;
X = [_A,_B],
Z = [_A,_B|Y] ? ;
X = [_A,_B,_C],
Z = [_A,_B,_C|Y] ? ;
X = [_A,_B,_C,_D],
Z = [_A,_B,_C,_D|Y] ? ;
X = [_A,_B,_C,_D,_E],
Z = [_A,_B,_C,_D,_E|Y] ? ;y
and unfortunately with ur add_tail i get this result
| ?- add_tail(X,Y,Z).
Z = X,
Y = [] ? ;
X = [_A],
Z = [_A|Y] ? ;
X = [_A|_B],
Y = [],
Z = [_A|_B] ? ;
X = [_A,_B],
Z = [_A,_B|Y] ? ;
X = [_A,_B|_C],
Y = [],
Z = [_A,_B|_C] ?
X = [_A,_B,_C],
Z = [_A,_B,_C|Y] ? y
yes
after a simple modification in your add_tail code i obtained your expected result
code
% add_tail(A,[],A):-! . comment
add_tail([],H,H) :-!.
add_tail([A|B],H,[A|C]) :- add_tail(B,H,C).
test add_tail
| ?- add_tail(X,Y,Z).
Z = Y,
X = [] ? ;
X = [_A],
Z = [_A|Y] ? ;
X = [_A,_B],
Z = [_A,_B|Y] ? ;
X = [_A,_B,_C],
Z = [_A,_B,_C|Y] ? ;
X = [_A,_B,_C,_D],
Z = [_A,_B,_C,_D|Y] ? ;
X = [_A,_B,_C,_D,_E],
Z = [_A,_B,_C,_D,_E|Y] ? y
yes
finaly
i test ur invert predicate without modification
| ?- invert([_A,_B,_C],L).
L = [_C,_B,_A] ? ;
no
I hope this post help you to explain how the predicate done inside
enjoy

Related

Extract other tuple member in Prolog

I want to write a predicate that takes a member of a tuple, that tuple, and outputs the other member of the tuple. You may assume tuples always have 2 elements, and the supplied member is always present.
extract_from_tuple(_, [], _).
extract_from_tuple(X, [H|T], R) :-
( X \= H -> R is X, ! ; X = H -> extract_from_tuple(X, T, R) ).
I tried to implement a simple if-else statement syntax I found on How to write IF ELSE conditions in Prolog.
So for example,
extract_from_tuple(a, [b,a], R).
should output b in the result variable R
Same should be for other way around:
extract_from_tuple(a, [a,b], R).
Only this time it 'should' hit the else statement and recursively call the predicate with the other element of the supplying list.
I think this problem is very simple and can be solved using just unification:
extract(X, [X,Y], Y).
extract(Y, [X,Y], X).
Examples:
?- extract(a, [b,a], R).
R = b.
?- extract(a, [a,b], R).
R = b ;
false.
To avoid spurious choice point, you can code extract/3 as:
extract_deterministic(Item, [First,Second], Rest) :-
( Item = First
-> Rest = Second
; Rest = First ).
Notice, however, this last version is less general than the first one! For example:
?- extract(X, [a,b], R). % two answers!
X = a,
R = b ;
X = b,
R = a.
?- extract_deterministic(X, [a,b], R). % only one answer!
X = a,
R = b.

Include non-unique elements only

This question was asked but there are no answers: here. I read the comments and tried to implement in both ways, but there are more problems that I don't understand.
I first tried the easy way that doesn't keep original order:
list_repeated(L, Ds) :-
msort(L, S),
sorted_repeated(S, Ds).
sorted_repeated([], []).
sorted_repeated([X|Xs], Ds) :-
first(Xs, X, Ds).
first([], _, []).
first([X|Xs], X, [X|Ds]) :-
more(Xs, X, Ds).
first([X|Xs], Y, Ds) :-
dif(X, Y),
first(Xs, X, Ds).
more([], _, []).
more([X|Xs], X, Ds) :-
more(Xs, X, Ds).
more([X|Xs], Y, Ds) :-
dif(X, Y),
first(Xs, X, Ds).
Once the list is sorted without removing duplicates, using first and more I add the element to the second argument if it occurs at least twice and skip all consecutive copies of the element.
This is not working properly because if I have:
?- list_duplicates([b,a,a,a,b,b], Ds).
I get answer [a,b] instead of [b,a] and also I get ; false after the answer.
I also tried another way, but this doesn't work because the accumulator is immutable?
list_duplicates(L, Ds) :-
ld_acc(L, [], Ds).
ld_acc([], _, []).
ld_acc([X|Xs], Acc, Ds) :-
( memberchk(X, Acc)
-> Ds = [X|Ds0],
ld_acc(Xs, Acc, Ds0)
; Acc1 = [X|Acc],
ld_acc(Xs, Acc1, Ds)
).
This cannot work because when I check that an element is member of accumulator I remove only one occurrence of each element: if I have three times the same element in the first argument, I am left with two. If I could change the element in the accumulator then I could maybe put a counter on it? In the first version I used different states, first and more, but here I have to attach state to the elements of the accumulator, is that possible?
A plea for purity
When programming in Prolog, a major attraction is the generality we enjoy from pure relations.
This lets us use our code in multiple directions, and reason declaratively over our programs and answers.
You can enjoy these benefits if you keep your programs pure.
Possible solution
As always when describing lists, also consider using DCG notation. See dcg for more information.
For example, to describe the list of duplicates in a pure way, consider:
list_duplicates([]) --> [].
list_duplicates([L|Ls]) -->
list_duplicates_(Ls, L),
list_duplicates(Ls).
list_duplicates_([], _) --> [].
list_duplicates_([L0|Ls], L) -->
if_(L0=L, [L], []),
list_duplicates_(Ls, L).
This uses if_//3 to retain generality and determinism (if applicable).
Examples
Here are a few example queries and answers. We start with simple ground cases:
?- phrase(list_duplicates([a,b,c]), Ds).
Ds = [].
?- phrase(list_duplicates([a,b,a]), Ds).
Ds = [a].
Even the most impure version will be able to handle these situations correctly. So, slightly more interesting:
?- phrase(list_duplicates([a,b,X]), Ds).
X = a,
Ds = [a] ;
X = b,
Ds = [b] ;
Ds = [],
dif(X, b),
dif(X, a).
Pretty nice, isn't it? The last part says: Ds = [] is a solution if X is different from b and a. Note the pure relation dif/2 automatically appears in these residual goals and retains the relation's generality.
Here is an example with two variables:
?- phrase(list_duplicates([X,Y]), Ds).
X = Y,
Ds = [Y] ;
Ds = [],
dif(Y, X).
Finally, consider using iterative deepening to fairly enumerate answers for lists of arbitrary length:
?- length(Ls, _), phrase(list_duplicates(Ls), Ds).
Ls = Ds, Ds = [] ;
Ls = [_136],
Ds = [] ;
Ls = [_136, _136],
Ds = [_136] ;
Ls = [_156, _162],
Ds = [],
dif(_162, _156) ;
Ls = Ds, Ds = [_42, _42, _42] ;
Ls = [_174, _174, _186],
Ds = [_174],
dif(_186, _174) .
Multiple occurrences
Here is a version that handles arbitrary many occurrences of the same element in such a way that exactly a single occurrence is retained if (and only if) the element occurs at least twice:
list_duplicates(Ls, Ds) :-
phrase(list_duplicates(Ls, []), Ds).
list_duplicates([], _) --> [].
list_duplicates([L|Ls], Ds0) -->
list_duplicates_(Ls, L, Ds0, Ds),
list_duplicates(Ls, Ds).
list_duplicates_([], _, Ds, Ds) --> [].
list_duplicates_([L0|Ls], L, Ds0, Ds) -->
if_(L0=L, new_duplicate(L0, Ds0, Ds1), {Ds0 = Ds1}),
list_duplicates_(Ls, L, Ds1, Ds).
new_duplicate(E, Ds0, Ds) -->
new_duplicate_(Ds0, E, Ds0, Ds).
new_duplicate_([], E, Ds0, [E|Ds0]) --> [E].
new_duplicate_([L|Ls], E, Ds0, Ds) -->
if_(L=E,
{ Ds0 = Ds },
new_duplicate_(Ls, E, Ds0, Ds)).
The query shown by #fatalize in the comments now yields:
?- list_duplicates([a,a,a], Ls).
Ls = [a].
The other examples yield the same results. For instance:
?- list_duplicates([a,b,c], Ds).
Ds = [].
?- list_duplicates([a,b,a], Ds).
Ds = [a].
?- list_duplicates([a,b,X], Ds).
X = a,
Ds = [a] ;
X = b,
Ds = [b] ;
Ds = [],
dif(X, b),
dif(X, a).
?- list_duplicates([X,Y], Ds).
X = Y,
Ds = [Y] ;
Ds = [],
dif(Y, X).
I leave the case ?- list_duplicates(Ls, Ls). as an exercise.
Generality: Multiple directions
Ideally, we want to be able to use a relation in all directions.
For example, our program should be able to answer questions like:
What does a list look like if its duplicates are [a,b]?
With the version shown above, we get:
?- list_duplicates(Ls, [a,b]).
nontermination
Luckily, a very simple change allows as to answer such questions!
One such change is to simply write:
list_duplicates(Ls, Ds) :-
length(Ls, _),
phrase(list_duplicates(Ls, []), Ds).
This is obviously declaratively admissible, because Ls must be a list. Operationally, this helps us to enumerate lists in a fair way.
We now get:
?- list_duplicates(Ls, [a,b]).
Ls = [a, a, b, b] ;
Ls = [a, b, a, b] ;
Ls = [a, b, b, a] ;
Ls = [a, a, a, b, b] ;
Ls = [a, a, b, a, b] ;
Ls = [a, a, b, b, a] ;
Ls = [a, a, b, b, b] ;
Ls = [a, a, b, b, _4632],
dif(_4632, b),
dif(_4632, a) ;
etc.
Here is a simpler case, using only a single element:
?- list_duplicates(Ls, [a]).
Ls = [a, a] ;
Ls = [a, a, a] ;
Ls = [a, a, _3818],
dif(_3818, a) ;
Ls = [a, _3870, a],
dif(_3870, a) ;
Ls = [_4058, a, a],
dif(a, _4058),
dif(a, _4058) ;
Ls = [a, a, a, a] ;
etc.
Maybe even more interesting:
What does a list without duplicates look like?
Our program answers:
?- list_duplicates(Ls, []).
Ls = [] ;
Ls = [_3240] ;
Ls = [_3758, _3764],
dif(_3764, _3758) ;
Ls = [_4164, _4170, _4176],
dif(_4176, _4164),
dif(_4176, _4170),
dif(_4170, _4164) .
Thus, the special case of a list where all elements are distinct naturally exists as a special case of the more general relation we have implemented.
We can use this relation to generate answers (as shown above), and also to test whether a list consists of distinct elements:
?- list_duplicates([a,b,c], []).
true.
?- list_duplicates([b,b], []).
false.
Unfortunately, the following even more general query still yields:
?- list_duplicates([b,b|_], []).
nontermination
On the plus side, if the length of the list is fixed, we get in such cases:
?- length(Ls, L), maplist(=(b), Ls),
( true ; list_duplicates(Ls, []) ).
Ls = [],
L = 0 ;
Ls = [],
L = 0 ;
Ls = [b],
L = 1 ;
Ls = [b],
L = 1 ;
Ls = [b, b],
L = 2 ;
Ls = [b, b, b],
L = 3 ;
Ls = [b, b, b, b],
L = 4 .
This is some indication that the program indeed terminates in such cases. Note that the answers are of course now too general.
Efficiency
It is well known in high-performance computing circles that as long as your program is fast enough, its correctness is barely worth considering.
So, the key question is of course: How can we make this faster?
I leave this is a very easy exercise. One way to make this faster in specific cases is to first check whether the given list is sufficiently instantiated. In that case, you can apply an ad hoc solution that fails terribly in more general cases, but has the extreme benefit that it is fast!
So as far as I can tell, you were on the right track with the accumulator, but this implementation definitely works as you want (assuming you want the duplicates in the order they first appear in the list).
list_duplicates(Input,Output) is just used to wrap and initialise the accumulator.
list_duplicates(Accumulator,[],Accumulator) unifies the accumulator with the output when we have finished processing the input list.
list_duplicates(Accumulator,[H|T],Output) says "if the head (H) of the input list is in the rest of the list (T), and is not in the Accumulator already, put it at the end of the Accumulator (using append), then recurse on the tail of the list".
list_duplicates(Accumulator,[_|T],Output) (which we only get to if either the head is not a duplicate, or is already in the Accumulator) just recurses on the tail of the list.
list_duplicates(Input,Output) :-
once(list_duplicates([],Input,Output)).
list_duplicates(Accumulator,[],Accumulator).
list_duplicates(Accumulator,[H|T],Output) :-
member(H,T),
\+member(H,Accumulator),
append(Accumulator,[H],NewAccumulator),
list_duplicates(NewAccumulator,T,Output).
list_duplicates(Accumulator,[_|T],Output) :-
list_duplicates(Accumulator,T,Output).
You could also recurse in list_duplicates(Accumulator,[H|T],Output) with list_duplicates([H|Accumulator],T,Output) and reverse in the wrapper, looking like this:
list_duplicates(Input,Output) :-
once(list_duplicates([],Input,ReverseOutput)),
reverse(ReverseOutput,Output).
list_duplicates(Accumulator,[],Accumulator).
list_duplicates(Accumulator,[H|T],Output) :-
member(H,T),
\+member(H,Accumulator),
list_duplicates([H|Accumulator],T,Output).
list_duplicates(Accumulator,[_|T],Output) :-
list_duplicates(Accumulator,T,Output).
The once call in the wrapper prevents the false output (or in this case, partial duplicate lists due to a lack of guards on the second rule).

Numbers in a list smaller than a given number

xMenores(_,[],[]).
xMenores(X,[H|T],[R|Z]) :-
xMenores(X,T,Z),
X > H,
R is H.
xMenores takes three parameters:
The first one is a number.
The second is a list of numbers.
The third is a list and is the variable that will contain the result.
The objective of the rule xMenores is obtain a list with the numbers of the list (Second parameter) that are smaller than the value on the first parameter. For example:
?- xMenores(3,[1,2,3],X).
X = [1,2]. % expected result
The problem is that xMenores returns false when X > H is false and my programming skills are almost null at prolog. So:
?- xMenores(4,[1,2,3],X).
X = [1,2,3]. % Perfect.
?- xMenores(2,[1,2,3],X).
false. % Wrong! "X = [1]" would be perfect.
I consider X > H, R is H. because I need that whenever X is bigger than H, R takes the value of H. But I don't know a control structure like an if or something in Prolog to handle this.
Please, any solution? Thanks.
Using ( if -> then ; else )
The control structure you might be looking for is ( if -> then ; else ).
Warning: you should probably swap the order of the first two arguments:
lessthan_if([], _, []).
lessthan_if([X|Xs], Y, Zs) :-
( X < Y
-> Zs = [X|Zs1]
; Zs = Zs1
),
lessthan_if(Xs, Y, Zs1).
However, if you are writing real code, you should almost certainly go with one of the predicates in library(apply), for example include/3, as suggested by #CapelliC:
?- include(>(3), [1,2,3], R).
R = [1, 2].
?- include(>(4), [1,2,3], R).
R = [1, 2, 3].
?- include(<(2), [1,2,3], R).
R = [3].
See the implementation of include/3 if you want to know how this kind of problems are solved. You will notice that lessthan/3 above is nothing but a specialization of the more general include/3 in library(apply): include/3 will reorder the arguments and use the ( if -> then ; else ).
"Declarative" solution
Alternatively, a less "procedural" and more "declarative" predicate:
lessthan_decl([], _, []).
lessthan_decl([X|Xs], Y, [X|Zs]) :- X < Y,
lessthan_decl(Xs, Y, Zs).
lessthan_decl([X|Xs], Y, Zs) :- X >= Y,
lessthan_decl(Xs, Y, Zs).
(lessthan_if/3 and lessthan_decl/3 are nearly identical to the solutions by Nicholas Carey, except for the order of arguments.)
On the downside, lessthan_decl/3 leaves behind choice points. However, it is a good starting point for a general, readable solution. We need two code transformations:
Replace the arithmetic comparisons < and >= with CLP(FD) constraints: #< and #>=;
Use a DCG rule to get rid of arguments in the definition.
You will arrive at the solution by lurker.
A different approach
The most general comparison predicate in Prolog is compare/3. A common pattern using it is to explicitly enumerate the three possible values for Order:
lessthan_compare([], _, []).
lessthan_compare([H|T], X, R) :-
compare(Order, H, X),
lessthan_compare_1(Order, H, T, X, R).
lessthan_compare_1(<, H, T, X, [H|R]) :-
lessthan_compare(T, X, R).
lessthan_compare_1(=, _, T, X, R) :-
lessthan_compare(T, X, R).
lessthan_compare_1(>, _, T, X, R) :-
lessthan_compare(T, X, R).
(Compared to any of the other solutions, this one would work with any terms, not just integers or arithmetic expressions.)
Replacing compare/3 with zcompare/3:
:- use_module(library(clpfd)).
lessthan_clpfd([], _, []).
lessthan_clpfd([H|T], X, R) :-
zcompare(ZOrder, H, X),
lessthan_clpfd_1(ZOrder, H, T, X, R).
lessthan_clpfd_1(<, H, T, X, [H|R]) :-
lessthan_clpfd(T, X, R).
lessthan_clpfd_1(=, _, T, X, R) :-
lessthan_clpfd(T, X, R).
lessthan_clpfd_1(>, _, T, X, R) :-
lessthan_clpfd(T, X, R).
This is definitely more code than any of the other solutions, but it does not leave behind unnecessary choice points:
?- lessthan_clpfd(3, [1,3,2], Xs).
Xs = [1, 2]. % no dangling choice points!
In the other cases, it behaves just as the DCG solution by lurker:
?- lessthan_clpfd(X, [1,3,2], Xs).
Xs = [1, 3, 2],
X in 4..sup ;
X = 3,
Xs = [1, 2] ;
X = 2,
Xs = [1] ;
X = 1,
Xs = [] .
?- lessthan_clpfd(X, [1,3,2], Xs), X = 3. %
X = 3,
Xs = [1, 2] ; % no error!
false.
?- lessthan_clpfd([1,3,2], X, R), R = [1, 2].
X = 3,
R = [1, 2] ;
false.
Unless you need such a general approach, include(>(X), List, Result) is good enough.
This can also be done using a DCG:
less_than([], _) --> [].
less_than([H|T], N) --> [H], { H #< N }, less_than(T, N).
less_than(L, N) --> [H], { H #>= N }, less_than(L, N).
| ?- phrase(less_than(R, 4), [1,2,3,4,5,6]).
R = [1,2,3] ? ;
You can write your predicate as:
xMenores(N, NumberList, Result) :- phrase(less_than(Result, N), NumberList).
You could write it as a one-liner using findall\3:
filter( N , Xs , Zs ) :- findall( X, ( member(X,Xs), X < N ) , Zs ) .
However, I suspect that the point of the exercise is to learn about recursion, so something like this would work:
filter( _ , [] , [] ) .
filter( N , [X|Xs] , [X|Zs] ) :- X < N , filter(N,Xs,Zs) .
filter( N , [X|Xs] , Zs ) :- X >= N , filter(N,Xs,Zs) .
It does, however, unpack the list twice on backtracking. An optimization here would be to combine the 2nd and 3rd clauses by introducing a soft cut like so:
filter( _ , [] , [] ) .
filter( N , [X|Xs] , [X|Zs] ) :-
( X < N -> Zs = [X|Z1] ; Zs = Z1 ) ,
filter(N,Xs,Zs)
.
(This is more like a comment than an answer, but too long for a comment.)
Some previous answers and comments have suggested using "if-then-else" (->)/2 or using library(apply) meta-predicate include/3. Both methods work alright, as long as only plain-old Prolog arithmetics—is/2, (>)/2, and the like—are used ...
?- X = 3, include(>(X),[1,3,2,5,4],Xs).
X = 3, Xs = [1,2].
?- include(>(X),[1,3,2,5,4],Xs), X = 3.
ERROR: >/2: Arguments are not sufficiently instantiated
% This is OK. When instantiation is insufficient, an exception is raised.
..., but when doing the seemingly benign switch from (>)/2 to (#>)/2, we lose soundness!
?- X = 3, include(#>(X),[1,3,2,5,4],Xs).
X = 3, Xs = [1,2].
?- include(#>(X),[1,3,2,5,4],Xs), X = 3.
false.
% This is BAD! Expected success with answer substitutions `X = 3, Xs = [1,2]`.
No new code is presented in this answer.
In the following we take a detailed look at different revisions of this answer by #lurker.
Revision #1, renamed to less_than_ver1//2. By using dcg and clpfd, the code is both very readable and versatile:
less_than_ver1(_, []) --> [].
less_than_ver1(N, [H|T]) --> [H], { H #< N }, less_than_ver1(N, T).
less_than_ver1(N, L) --> [H], { H #>= N }, less_than_ver1(N, L).
Let's query!
?- phrase(less_than_ver1(N,Zs),[1,2,3,4,5]).
N in 6..sup, Zs = [1,2,3,4,5]
; N = 5 , Zs = [1,2,3,4]
; N = 4 , Zs = [1,2,3]
; N = 3 , Zs = [1,2]
; N = 2 , Zs = [1]
; N in inf..1, Zs = []
; false.
?- N = 3, phrase(less_than_ver1(N,Zs),[1,2,3,4,5]).
N = 3, Zs = [1,2] % succeeds, but leaves useless choicepoint
; false.
?- phrase(less_than_ver1(N,Zs),[1,2,3,4,5]), N = 3.
N = 3, Zs = [1,2]
; false.
As a small imperfection, less_than_ver1//2 leaves some useless choicepoints.
Let's see how things went with the newer revision...
Revision #3, renamed to less_than_ver3//2:
less_than_ver3([],_) --> [].
less_than_ver3(L,N) --> [X], { X #< N -> L=[X|T] ; L=T }, less_than_ver3(L,N).
This code uses the if-then-else ((->)/2 + (;)/2) in order to improve determinism.
Let's simply re-run the above queries!
?- phrase(less_than_ver3(Zs,N),[1,2,3,4,5]).
N in 6..sup, Zs = [1,2,3,4,5]
; false. % all other solutions are missing!
?- N = 3, phrase(less_than_ver3(Zs,N),[1,2,3,4,5]).
N = 3, Zs = [1,2] % works as before, but no better.
; false. % we still got the useless choicepoint
?- phrase(less_than_ver3(Zs,N),[1,2,3,4,5]), N = 3.
false. % no solution!
% we got one with revision #1!
Surprise! Two cases that worked before are now (somewhat) broken, and the determinism in the ground case is no better... Why?
The vanilla if-then-else often cuts too much too soon, which is particularly problematic with code which uses coroutining and/or constraints.
Note that (*->)/2 (a.k.a. "soft-cut" or if/3), fares only a bit better, not a lot!
As if_/3 never ever cuts more (often than) the vanilla if-then-else (->)/2, it cannot be used in above code to improve determinism.
If you want to use if_/3 in combination with constraints, take a step back and write code that is non-dcg as the first shot.
If you're lazy like me, consider using a meta-predicate like tfilter/3 and (#>)/3.
This answer by #Boris presented a logically pure solution which utilizes clpfd:zcompare/3 to help improve determinism in certain (ground) cases.
In this answer we will explore different ways of coding logically pure Prolog while trying to avoid the creation of useless choicepoints.
Let's get started with zcompare/3 and (#<)/3!
zcompare/3 implements three-way comparison of finite domain variables and reifies the trichotomy into one of <, =, or >.
As the inclusion criterion used by the OP was a arithmetic less-than test, we propose using
(#<)/3 for reifying the dichotomy into one of true or false.
Consider the answers of the following queries:
?- zcompare(Ord,1,5), #<(1,5,B).
Ord = (<), B = true.
?- zcompare(Ord,5,5), #<(5,5,B).
Ord = (=), B = false.
?- zcompare(Ord,9,5), #<(9,5,B).
Ord = (>), B = false.
Note that for all items to be selected both Ord = (<) and B = true holds.
Here's a side-by-side comparison of three non-dcg solutions based on clpfd:
The left one uses zcompare/3 and first-argument indexing on the three cases <, =, and >.
The middle one uses (#<)/3 and first-argument indexing on the two cases true and false.
The right one uses (#<)/3 in combination with if_/3.
Note that we do not need to define auxiliary predicates in the right column!
less_than([],[],_). % less_than([],[],_). % less_than([],[],_).
less_than([Z|Zs],Ls,X) :- % less_than([Z|Zs],Ls,X) :- % less_than([Z|Zs],Ls,X) :-
zcompare(Ord,Z,X), % #<(Z,X,B), % if_(Z #< X,
ord_lt_(Ord,Z,Ls,Rs), % incl_lt_(B,Z,Ls,Rs), % Ls = [Z|Rs],
less_than(Zs,Rs,X). % less_than(Zs,Rs,X). % Ls = Rs),
% % less_than(Zs,Rs,X).
ord_lt_(<,Z,[Z|Ls],Ls). % incl_lt_(true ,Z,[Z|Ls],Ls). %
ord_lt_(=,_, Ls ,Ls). % incl_lt_(false,_, Ls ,Ls). %
ord_lt_(>,_, Ls ,Ls). % %
Next, let's use dcg!
In the right column we use if_//3 instead of if_/3.
Note the different argument orders of dcg and non-dcg solutions: less_than([1,2,3],Zs,3) vs phrase(less_than([1,2,3],3),Zs).
The following dcg implementations correspond to above non-dcg codes:
less_than([],_) --> []. % less_than([],_) --> []. % less_than([],_) --> [].
less_than([Z|Zs],X) --> % less_than([Z|Zs],X) --> % less_than([Z|Zs],X) -->
{ zcompare(Ord,Z,X) }, % { #<(Z,X,B) }, % if_(Z #< X,[Z],[]),
ord_lt_(Ord,Z), % incl_lt_(B,Z), % less_than(Zs,X).
less_than(Zs,X). % less_than(Zs,X). %
% %
ord_lt_(<,Z) --> [Z]. % incl_lt_(true ,Z) --> [Z]. %
ord_lt_(=,_) --> []. % incl_lt_(false,_) --> []. %
ord_lt_(>,_) --> []. % %
OK! Saving the best for last... Simply use meta-predicate tfilter/3 together with (#>)/3!
less_than(Xs,Zs,P) :-
tfilter(#>(P),Xs,Zs).
The dcg variant in this previous answer is our starting point.
Consider the auxiliary non-terminal ord_lt_//2:
ord_lt_(<,Z) --> [Z].
ord_lt_(=,_) --> [].
ord_lt_(>,_) --> [].
These three clauses can be covered using two conditions:
Ord = (<): the item should be included.
dif(Ord, (<)): it should not be included.
We can express this "either-or choice" using if_//3:
less_than([],_) --> [].
less_than([Z|Zs],X) -->
{ zcompare(Ord,Z,X) },
if_(Ord = (<), [Z], []),
less_than(Zs,X).
Thus ord_lt_//2 becomes redundant.
Net gain? 3 lines-of-code !-)

Element appears exactly once in the list in Prolog

I want to write a predicate which check if the Element appeares exactly once in the List.
once(Element, List).
My code:
once(X, [H | T]) :-
\+ X = H,
once(X, T).
once(X, [X | T]) :-
\+ member(X, T).
?- once(c, [b,a,a,c,b,a]).
true
?- once(b, [b,a,a,c,b,a]).
false.
But if I ask:
once(X, [b,a,a,c,b,a]).
Prolog answers:
false
Why? Prolog should find X = c solution. Where is bug?
Running a trace in prolog can be very helpful in determining the answer to this sort of question. We'll do the trace manually here for illustration.
Let's look at your predicate:
once(X, [H | T]) :-
\+ X = H,
once(X, T).
once(X, [X | T]) :-
\+ member(X, T).
Let's consider now the query:
once(X, [b,a,a,c,b,a]).
First, Prolog attempts the first clause of your predicate. The head is once(X, [H|T]) and first expression is \+ X = H, which will become:
once(X, [b|[a,a,c,b,a]]) :- % [H|T] instantiated with [b,a,a,c,b,a] here
% So, H is b, and T is [a,a,c,b,a]
\+ X = b,
...
X is instantiated (be unified with) with the atom b here, and the result of that unification succeeds. However, you have a negation in front of this, so the result of \+ X = b, when X is initially unbound, will be false since X = b unifies X with b and is true.
The first clause thus fails. Prolog moves to the next clause. The clause head is once(X, [X|T]) and following is \+ member(X, T), which become:
once(b, [b|[a,a,c,b,a]]) :- % X was instantiated with 'b' here,
% and T instantiated with [a,a,c,b,a]
\+ member(b, [a,a,c,b,a]).
member(b, [a,a,c,b,a]) succeeds because b is a member of [a,a,c,b,a]. Therefore, \+ member(b, [a,a,c,b,a]) fails.
The second clause fails, too.
There are no more clauses for the predicate once(X, [b,a,a,c,b,a]). All of them failed. So the query fails. The primary issue is that \+ X = H (or even X \= H, when X is not instantiated, won't choose a value from the list that is not the same as the value instantiated in H. Its behavior isn't logically what you want.
A more straight-on approach to the predicate would be:
once(X, L) :- % X occurs once in L if...
select(X, L, R), % I can remove X from L giving R, and
\+ member(X, R). % X is not a member of R
The select will query as desired for uninstantiated X, so this will yield:
?- once(c, [b,a,a,c,b,a]).
true ;
false.
?- once(b, [b,a,a,c,b,a]).
false.
?- once(X, [b,a,a,c,b,a]).
X = c ;
false.
As an aside, I'd avoid the predicate name once since it is the name of a built-in predicate in Prolog. But it has no bearing on this particular problem.
Using prolog-dif we can preserve logical-purity!
The following code is based on the previous answer by #lurker, but is logically pure:
onceMember_of(X,Xs) :-
select(X,Xs,Xs0),
maplist(dif(X),Xs0).
Let's look at some queries:
?- onceMember_of(c,[b,a,a,c,b,a]).
true ; % succeeds, but leaves choicepoint
false.
?- onceMember_of(b,[b,a,a,c,b,a]).
false.
?- onceMember_of(X,[b,a,a,c,b,a]).
X = c ;
false.
The code is monotone, so we get logically sound answers for more general uses, too!
?- onceMember_of(X,[A,B,C]).
X = A, dif(A,C), dif(A,B) ;
X = B, dif(B,C), dif(B,A) ;
X = C, dif(C,B), dif(C,A) ;
false.
Let's look at all lists in increasing size:
?- length(Xs,_), onceMember_of(X,Xs).
Xs = [X] ;
Xs = [X,_A], dif(X,_A) ;
Xs = [_A,X], dif(X,_A) ;
Xs = [ X,_A,_B], dif(X,_A), dif(X,_B) ;
Xs = [_A, X,_B], dif(X,_A), dif(X,_B) ;
Xs = [_A,_B, X], dif(X,_A), dif(X,_B) ;
Xs = [ X,_A,_B,_C], dif(X,_A), dif(X,_B), dif(X,_C) ...
At last, let's have the most general query:
?- onceMember_of(X,Xs).
Xs = [X] ;
Xs = [X,_A], dif(X,_A) ;
Xs = [X,_A,_B], dif(X,_A), dif(X,_B) ;
Xs = [X,_A,_B,_C], dif(X,_A), dif(X,_B),dif(X,_C) ...
Edit 2015-05-13
We can do even better by using selectfirst/3, a drop-in replacement of select/3:
onceMember_ofB(X,Xs) :-
selectfirst(X,Xs,Xs0),
maplist(dif(X),Xs0).
Let's run onceMember_of/2 and onceMember_ofB/2 head to head:
?- onceMember_of(c,[b,a,a,c,b,a]).
true ; % succeeds, but leaves choicepoint
false.
?- onceMember_ofB(c,[b,a,a,c,b,a]).
true. % succeeds deterministically
But we can still get better! Consider:
?- onceMember_ofB(X,[A,B,C]).
X = A, dif(A,C), dif(A,B) ;
X = B, dif(B,C), dif(A,B),dif(B,A) ; % 1 redundant constraint
X = C, dif(A,C),dif(C,A), dif(B,C),dif(C,B) ; % 2 redundant constraints
false.
Note the redundant dif/2 constraints? They come from the goal
maplist(dif(X),Xs0) and we can eliminate them, like so:
onceMember_ofC(E,[X|Xs]) :-
if_(E = X, maplist(dif(X),Xs),
onceMember_ofC(E,Xs)).
Let's see if it works!
?- onceMember_ofC(X,[A,B,C]).
X = A, dif(A,C), dif(A,B) ;
X = B, dif(B,C), dif(B,A) ;
X = C, dif(C,B), dif(C,A) ;
false.

Add two more occurrences using prolog

I have a list [a, b, a, a, a, c, c]
and I need to add two more occurrences of each element.
The end result should look like this:
[a, a, a, b, b, b, a, a, a, a, a, c, c, c, c]
If I have an item on the list that is the same as the next item, then it keeps going until there is a new item, when it finds the new item, it adds two occurrences of the previous item then moves on.
This is my code so far, but I can't figure out how to add two...
dbl([], []).
dbl([X], [X,X]).
dbl([H|T], [H,H|T], [H,H|R]) :- dbl(T, R).
Your code looks a bit strange because the last rule takes three parameters. You only call the binary version, so no recursion will ever try to derive it.
You already had a good idea to look at the parts of the list, where elements change. So there are 4 cases:
1) Your list is empty.
2) You have exactly one element.
3) Your list starts with two equal elements.
4) Your list starts with two different elements.
Case 1 is not specified, so you might need to find a sensible choice for that. Case 2 is somehow similar to case 4, since the end of the list can be seen as a change in elements, where you need to append two copies, but then you are done. Case 3 is quite simple, we can just keep the element and recurse on the rest. Case 4 is where you need to insert the two copies again.
This means your code will look something like this:
% Case 1
dbl([],[]).
% Case 2
dbl([X],[X,X,X]).
% Case 3
dbl([X,X|Xs], [X|Ys]) :-
% [...] recursion skipping the leading X
% Case 4
dbl([X,Y|Xs], [X,X,X|Ys]) :-
dif(X,Y),
% [...] we inserted the copies, so recursion on [Y|Xs] and Ys
Case 3 should be easy to finish, we just drop the first X from both lists and recurse on dbl([X|Xs],Ys). Note that we implicitly made the first two elements equal (i.e. we unified them) by writing the same variable twice.
If you look at the head of case 4, you can directly imitate the pattern you described: supposed the list starts with X, then Y and they are different (dif(X,Y)), the X is repeated 3 times instead of just copied and we then continue with the recursion on the rest starting with Y: dbl([Y|Xs],Ys).
So let's try out the predicate:
?- dbl([a,b,a,a,a,c,c],[a,a,a,b,b,b,a,a,a,a,a,c,c,c,c]).
true ;
false.
Our test case is accepted (true) and we don't find more than one solution (false).
Let's see if we find a wrong solution:
?- dif(Xs,[a,a,a,b,b,b,a,a,a,a,a,c,c,c,c]), dbl([a,b,a,a,a,c,c],Xs).
false.
No, that's also good. What happens, if we have variables in our list?
?- dbl([a,X,a],Ys).
X = a,
Ys = [a, a, a, a, a] ;
Ys = [a, a, a, X, X, X, a, a, a],
dif(X, a),
dif(X, a) ;
false.
Either X = a, then Ys is single run of 5 as; or X is not equal to a, then we need to append the copies in all three runs. Looks also fine. (*)
Now lets see, what happens if we only specify the solution:
?- dbl(X,[a,a,a,b,b]).
false.
Right, a list with a run of only two bs can not be a result of our specification. So lets try to add one:
?- dbl(X,[a,a,a,b,b,b]).
X = [a, b] ;
false.
Hooray, it worked! So lets as a last test look what happens, if we just call our predicate with two variables:
?- dbl(Xs,Ys).
Xs = Ys, Ys = [] ;
Xs = [_G15],
Ys = [_G15, _G15, _G15] ;
Xs = [_G15, _G15],
Ys = [_G15, _G15, _G15, _G15] ;
Xs = [_G15, _G15, _G15],
Ys = [_G15, _G15, _G15, _G15, _G15] ;
Xs = [_G15, _G15, _G15, _G15],
Ys = [_G15, _G15, _G15, _G15, _G15, _G15] ;
[...]
It seems we get the correct answers, but we see only cases for a single run. This is a result of prolog's search strategy(which i will not explain in here). But if we look at shorter lists before we generate longer ones, we can see all the solutions:
?- length(Xs,_), dbl(Xs,Ys).
Xs = Ys, Ys = [] ;
Xs = [_G16],
Ys = [_G16, _G16, _G16] ;
Xs = [_G16, _G16],
Ys = [_G16, _G16, _G16, _G16] ;
Xs = [_G86, _G89],
Ys = [_G86, _G86, _G86, _G89, _G89, _G89],
dif(_G86, _G89) ;
Xs = [_G16, _G16, _G16],
Ys = [_G16, _G16, _G16, _G16, _G16] ;
Xs = [_G188, _G188, _G194],
Ys = [_G188, _G188, _G188, _G188, _G194, _G194, _G194],
dif(_G188, _G194) ;
[...]
So it seems we have a working predicate (**), supposed you filled in the missing goals from the text :)
(*) A remark here: this case only works because we are using dif. The first predicates with equality, one usually encounters are =, == and their respective negations \= and \==. The = stands for unifyability (substituting variables in the arguments s.t. they become equal) and the == stands for syntactic equality (terms being exactly equal). E.g.:
?- f(X) = f(a).
X = a.
?- f(X) \= f(a).
false.
?- f(X) == f(a).
false.
?- f(X) \== f(a).
true.
This means, we can make f(X) equal to f(a), if we substitute X by a. This means if we ask if they can not be made equal (\=), we get the answer false. On the other hand, the two terms are not equal, so == returns false, and its negation \== answers true.
What this also means is that X \== Y is always true, so we can not use \== in our code. In contrast to that, dif waits until it can decide wether its arguments are equal or not. If this is still undecided after finding an answer, the "dif(X,a)" statements are printed.
(**) One last remark here: There is also a solution with the if-then-else construct (test -> goals_if_true; goals_if_false, which merges cases 3 and 4. Since i prefer this solution, you might need to look into the other version yourself.
TL;DR:
From a declarative point of view, the code sketched by #lambda.xy.x is perfect.
Its determinacy can be improved without sacrificing logical-purity.
Code variant #0: #lambda.xy.x's code
Here's the code we want to improve:
dbl0([], []).
dbl0([X], [X,X,X]).
dbl0([X,X|Xs], [X|Ys]) :-
dbl0([X|Xs], Ys).
dbl0([X,Y|Xs], [X,X,X|Ys]) :-
dif(X, Y),
dbl0([Y|Xs], Ys).
Consider the following query and the answer SWI-Prolog gives us:
?- dbl0([a],Xs).
Xs = [a,a,a] ;
false.
With ; false the SWI prolog-toplevel
indicates a choicepoint was left when proving the goal.
For the first answer, Prolog did not search the entire proof tree.
Instead, it replied "here's an answer, there may be more".
Then, when asked for more solutions, Prolog traversed the remaining branches of the proof tree but finds no more answers.
In other words: Prolog needs to think twice to prove something we knew all along!
So, how can we give determinacy hints to Prolog?
By utilizing:
control constructs !/0 and / or (->)/2 (potentially impure)
first argument indexing on the principal functor (never impure)
The code presented in the earlier answer by #CapelliC—which is based on !/0, (->)/2, and the meta-logical predicate (\=)/2—runs well if all arguments are sufficiently instantiated. If not, erratic answers may result—as #lambda.xy.x's comment shows.
Code variant #1: indexing
Indexing can improve determinacy without ever rendering the code non-monotonic. While different Prolog processors have distinct advanced indexing capabilities, the "first-argument principal-functor" indexing variant is widely available.
Principal? This is why executing the goal dbl0([a],Xs) leaves a choicepoint behind: Yes, the goal only matches one clause—dbl0([X],[X,X,X]).—but looking no deeper than the principal functor Prolog assumes that any of the last three clauses could eventually get used. Of course, we know better...
To tell Prolog we utilize principal-functor first-argument indexing:
dbl1([], []).
dbl1([E|Es], Xs) :-
dbl1_(Es, Xs, E).
dbl1_([], [E,E,E], E).
dbl1_([E|Es], [E|Xs], E) :-
dbl1_(Es, Xs, E).
dbl1_([E|Es], [E0,E0,E0|Xs], E0) :-
dif(E0, E),
dbl1_(Es, Xs, E).
Better? Somewhat, but determinacy could be better still...
Code variant #2: indexing on reified term equality
To make Prolog see that the two recursive clauses of dbl1_/3 are mutually exclusive (in certain cases), we reify the truth value of
term equality and then index on that value:
This is where reified term equality (=)/3 comes into play:
dbl2([], []).
dbl2([E|Es], Xs) :-
dbl2_(Es, Xs, E).
dbl2_([], [E,E,E], E).
dbl2_([E|Es], Xs, E0) :-
=(E0, E, T),
t_dbl2_(T, Xs, E0, E, Es).
t_dbl2_(true, [E|Xs], _, E, Es) :-
dbl2_(Es, Xs, E).
t_dbl2_(false, [E0,E0,E0|Xs], E0, E, Es) :-
dbl2_(Es, Xs, E).
Sample queries using SWI-Prolog:
?- dbl0([a],Xs).
Xs = [a, a, a] ;
false.
?- dbl1([a],Xs).
Xs = [a, a, a].
?- dbl2([a],Xs).
Xs = [a, a, a].
?- dbl0([a,b,b],Xs).
Xs = [a, a, a, b, b, b, b] ;
false.
?- dbl1([a,b,b],Xs).
Xs = [a, a, a, b, b, b, b] ;
false.
?- dbl2([a,b,b],Xs).
Xs = [a, a, a, b, b, b, b].
To make above code more compact, use control construct if_/3 .
I was just about to throw this version with if_/3 and (=)/3 in the hat when I saw #repeat already suggested it. So this is essentially the more compact version as outlined by #repeat:
list_dbl([],[]).
list_dbl([X],[X,X,X]).
list_dbl([A,B|Xs],DBL) :-
if_(A=B,DBL=[A,B|Ys],DBL=[A,A,A,B|Ys]),
list_dbl([B|Xs],[B|Ys]).
It yields the same results as dbl2/2 by #repeat:
?- list_dbl([a],DBL).
DBL = [a,a,a]
?- list_dbl([a,b,b],DBL).
DBL = [a,a,a,b,b,b,b]
The example query by the OP works as expected:
?- list_dbl([a,b,a,a,a,c,c],DBL).
DBL = [a,a,a,b,b,b,a,a,a,a,a,c,c,c,c]
Plus here are some of the example queries provided by #lambda.xy.x. They yield the same results as #repeat's dbl/2 and #lambda.xy.x's dbl/2:
?- dif(Xs,[a,a,a,b,b,b,a,a,a,a,a,c,c,c,c]), list_dbl([a,b,a,a,a,c,c],Xs).
no
?- list_dbl(X,[a,a,a,b,b]).
no
?- list_dbl(L,[a,a,a,b,b,b]).
L = [a,b] ? ;
no
?- list_dbl(L,DBL).
DBL = L = [] ? ;
DBL = [_A,_A,_A],
L = [_A] ? ;
DBL = [_A,_A,_A,_A],
L = [_A,_A] ? ;
DBL = [_A,_A,_A,_A,_A],
L = [_A,_A,_A] ? ;
...
?- list_dbl([a,X,a],DBL).
DBL = [a,a,a,a,a],
X = a ? ;
DBL = [a,a,a,X,X,X,a,a,a],
dif(X,a),
dif(a,X)
?- length(L,_), list_dbl(L,DBL).
DBL = L = [] ? ;
DBL = [_A,_A,_A],
L = [_A] ? ;
DBL = [_A,_A,_A,_A],
L = [_A,_A] ? ;
DBL = [_A,_A,_A,_B,_B,_B],
L = [_A,_B],
dif(_A,_B) ? ;
DBL = [_A,_A,_A,_A,_A],
L = [_A,_A,_A] ?
dbl([X,Y|T], [X,X,X|R]) :- X \= Y, !, dbl([Y|T], R).
dbl([H|T], R) :-
T = []
-> R = [H,H,H]
; R = [H|Q], dbl(T, Q).
The first clause handles the basic requirement, adding two elements on sequence change.
The second one handles list termination as a sequence change, otherwise, does a plain copy.