I’m using: SWI-Prolog versionSWI-Prolog version 9.2.9 for arm64-darwin. I want the code to: Solve problem number 8 from P-99: Ninety-Nine Prolog Problems : Eliminate consecutive duplicates of list elements. Questions: 1 - Isn’t the pattern matching depending on the order of clauses of a multi-cla…

If I remember correctly, Erlang is a committed choice language. If so there is an implicit cut/commit in the neck of every alternative. There is not an implicit cut in Prolog so the concept of “already covered” must be done by an explicit cut to remove the choicepoint when the first two elements un…

The task is to produce clumped/2 without the count - and (preferably) able to handle non-ground variables - which my code does at Logically pure/impure predicate implementations in the std lib - #21 by brebs Example: ?- clumped2([a,b,X,d], L). L = [a-1, b-1, X-1, d-1], dif(X, b), dif(X, d) ; X = d…

I don’t understand your first question, the heads of clauses of a predicate (like no_duplicates/2) are matched in sequence. Is this that worries you ? About the second and third questions, the cut will prevent the choicepoint required by the following clause, since inequality cannot be expressed in…

Awesome! Thanks! I don’t understand your first question, the heads of clauses of a predicate (like no_duplicates/2) are matched in sequence. Is this that worries you ? Yep! Because if that (being orderly) was the case the 4th clause no_duplicates([H1,H2|T], [H1|L1]) :- could simply be no_duplica…

[image] dc0d: The cut Example: Programming cooperation - #20 by brebs

[image] ridgeworks: Good advice. Agree. Note that on the toplevel you can also use the * command after a query succeeds with a choicepoint: ?- no_duplicates([1,1,1,2,2,2], X). X = [1, 2] <press * here> % no_duplicates([2],[2]) left a choice point in alternate clause (after success) % /ho…

(… in Erlang …) there is an implicit cut/commit in the neck of every alternative That was so to the point (in the world of my incomplete understanding of how to think in “relations”)! Managed to rewrite it with two cuts (and no conditions - and I like the fourth clause now): :- module(workbench…

A couple of additional points: You can unify in the head of a clause so the second clause can just be: no_duplicates([E], [E]) :- !. If you look at the doc “if-then-else” control predicate (written as Cond -> Goal1 ; Goal2), you’ll see that it does have an implicit cut. So you can replace the l…

[image] dc0d: no_duplicates([], []). no_duplicates([E], Out) :- !, Out=[E]. no_duplicates([H,H|T], Out) :- !, no_duplicates([H|T], Out). no_duplicates([H|T], [H|L1]) :- no_duplicates(T, L1). Note that the 2nd clause is unneeded as the list of one element is covered by the last and first claus…

[image] dc0d: no_duplicates([H,H|T], Out) :- !, no_duplicates([H|T], Out). Such code is flawed - it will only work as intended if its input variables are ground/1 Should be using dif/2 E.g. if given: no_duplicates([X, Y], L) … the output should be very similar to: L = [X, Y], dif(X, Y) ;…

[image] brebs: Such code is flawed Please consider that dif/2 was not deemed necessary in ‘Programming in Prolog’ by Clocksin-Mellish.

Pattern Matching Order

Help!

brebs February 17, 2025, 4:43pm 6

Example: Programming cooperation - #20 by brebs

Topic		Replies	Views
I don't understand why this query do not terminate Help!	24	1887	March 31, 2022
Why does this predicate change a list with a tail, to a list with two elements, one ground, one var? Help!	15	1943	July 17, 2019
Logically pure/impure predicate implementations in the std lib General	16	935	May 19, 2023
Coding problem Help!	23	2025	September 12, 2020
Another day, another predicate, another choicepoint! General	15	516	September 5, 2023

Pattern Matching Order

Related topics