# Copy\_term but without maintaining variable relationships? **URL:** \ **Category:** Predicate\ **Created:** [July 26, 2021, 5:25pm UTC](https://swi-prolog.discourse.group/t/copy-term-but-without-maintaining-variable-relationships/4204 "2021-07-26T17:25:00Z")\ **Posts on this page:** 11\ **Page:** 1
**Author:** ![dmchurch](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/dmchurch/32/3427_2.png) [@dmchurch](https://swi-prolog.discourse.group/u/dmchurch)\ **Post date:** [July 26, 2021, 5:25pm UTC](https://swi-prolog.discourse.group/t/copy-term-but-without-maintaining-variable-relationships/4204/1 "2021-07-26T17:25:00Z")
Is there any predicate for making a copy of a compound term and un-unifying any unified variables, so that the resulting term has only singletons? In other words, is there something like a copy\_to\_singletons/2 that would let me do this: ```prolog ?- copy_to_singletons(foo(A, A), X). X = foo(_1, _2). ``` I’ve been using this as a workaround, but it seems needlessly complex: ```prolog copy_to_singletons(Term, TermOut) :- copy_term(Term, Term1, _), numbervars(Term1), mapsubterms([T0,_]>>var_number(T0, _), Term1, TermOut), ``` I end up having to make two copies because for whatever reason, mapsubterms/3 doesn’t call the goal for variables. Not sure what the reasoning is behind that, but it does mean I have to copy and ground before calling it. Is there a better way? ---
**Author:** ![peter.ludemann](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/peter.ludemann/32/48_2.png) [@peter.ludemann](https://swi-prolog.discourse.group/u/peter.ludemann)\ **Post date:** [July 26, 2021, 10:23pm UTC](https://swi-prolog.discourse.group/t/copy-term-but-without-maintaining-variable-relationships/4204/2 "2021-07-26T22:23:24Z")
I’m curious why you want to do this. 😉 Anyway library(varnumbers) has varnumbers/3, which might do what you want instead of mapsubterms. ---
**Author:** ![dmchurch](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/dmchurch/32/3427_2.png) [@dmchurch](https://swi-prolog.discourse.group/u/dmchurch)\ **Post date:** [July 26, 2021, 11:22pm UTC](https://swi-prolog.discourse.group/t/copy-term-but-without-maintaining-variable-relationships/4204/3 "2021-07-26T23:22:13Z")
Unfortunately, varnumbers/3 is going to (by design) have the same problem as copy\_term, which is to say that common variables in the original will become common variables in the copy. As for what I’m trying to do, I want to reason about the structure of Prolog clauses themselves, so a non-singleton variable is _significant_ but that does not mean the two (or more) instances will be _identical_. ---
**Author:** ![jan](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/jan/32/4_2.png) [@jan](https://swi-prolog.discourse.group/u/jan)\ **Post date:** [July 27, 2021, 7:14am UTC](https://swi-prolog.discourse.group/t/copy-term-but-without-maintaining-variable-relationships/4204/4 "2021-07-27T07:14:19Z")
> [@dmchurch](#): > > I want to reason about the structure of Prolog clauses themselves, so a non-singleton variable is _significant_ but that does not mean the two (or more) instances will be _identical_ . I don’t understand that. I’ve seen a lot of things during my Prolog years, but never the desire to stop sharing two variables. As they are clause structures, they are not cyclic and thus a simple transformation as mapsubterms/3 does the job. You can’t use that as it doesn’t call the mapper for variables (it is still unclear whether that is a good decision. There are arguments both ways). ---
**Author:** ![dmchurch](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/dmchurch/32/3427_2.png) [@dmchurch](https://swi-prolog.discourse.group/u/dmchurch)\ **Post date:** [July 27, 2021, 7:33am UTC](https://swi-prolog.discourse.group/t/copy-term-but-without-maintaining-variable-relationships/4204/5 "2021-07-27T07:33:10Z")
> [@jan](#): > > I’ve seen a lot of things during my Prolog years, but never the desire to stop sharing two variables. Well, it’s always nice to be unique 😂 I want to map them into `term_position`s. I don’t want to use them semantically, I want to use them syntactically, and they are two different tokens in the stream. ---
**Author:** ![ridgeworks](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/ridgeworks/32/886_2.png) [@ridgeworks](https://swi-prolog.discourse.group/u/ridgeworks)\ **Post date:** [July 27, 2021, 1:27pm UTC](https://swi-prolog.discourse.group/t/copy-term-but-without-maintaining-variable-relationships/4204/6 "2021-07-27T13:27:50Z")
Regardless of motivation this looks like a problem with a straight forward Prolog solution. Something like: ```prolog copy_to_singletons(X, _) :- var(X), !. copy_to_singletons(X, X) :- atomic(X), !. copy_to_singletons([X|Xs], [S|Ss]) :- !, copy_to_singletons(X,S), copy_to_singletons(Xs,Ss). copy_to_singletons(X, S) :- compound(X), X =.. [F|Xs], copy_to_singletons(Xs,Ss), S =.. [F|Ss]. ``` so: ```prolog ?- copy_to_singletons(foo(A,A),X). X = foo(_4786,_4792). ``` Am I missing something? ---
**Author:** ![peter.ludemann](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/peter.ludemann/32/48_2.png) [@peter.ludemann](https://swi-prolog.discourse.group/u/peter.ludemann)\ **Post date:** [July 27, 2021, 3:33pm UTC](https://swi-prolog.discourse.group/t/copy-term-but-without-maintaining-variable-relationships/4204/8 "2021-07-27T15:33:51Z")
These lines aren’t needed (although they might speed things up): > [@ridgeworks](#): > > ```prolog > copy_to_singletons([X|Xs], [S|Ss]) :- !, > copy_to_singletons(X,S), > copy_to_singletons(Xs,Ss). > > ``` because: ```prolog ?- [a|b] =.. F. F = ['[|]', a, b]. ``` and instead of (=…)/2, probably better to use compound\_name\_arguments/3. Or mapargs/3 could be used: ```prolog copy_to_singletons(X, _) :- var(X), !. copy_to_singletons(X, X) :- atomic(X), !. copy_to_singletons(X, S) :- mapargs(copy_to_singletons, X, S). ``` ---
**Author:** ![ridgeworks](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/ridgeworks/32/886_2.png) [@ridgeworks](https://swi-prolog.discourse.group/u/ridgeworks)\ **Post date:** [July 27, 2021, 4:01pm UTC](https://swi-prolog.discourse.group/t/copy-term-but-without-maintaining-variable-relationships/4204/9 "2021-07-27T16:01:05Z")
> [@peter.ludemann](#): > > These lines aren’t needed (although they might speed things up): That was my first thought, but try it on the example. > [@peter.ludemann](#): > > Or [mapargs/3](https://www.swi-prolog.org/pldoc/doc_for?object=mapargs/3) could be used: Probably the best solution, but (like compound\_name\_arguments/3) not particularly portable. ---
**Author:** ![peter.ludemann](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/peter.ludemann/32/48_2.png) [@peter.ludemann](https://swi-prolog.discourse.group/u/peter.ludemann)\ **Post date:** [July 27, 2021, 4:14pm UTC](https://swi-prolog.discourse.group/t/copy-term-but-without-maintaining-variable-relationships/4204/10 "2021-07-27T16:14:12Z")
> [@ridgeworks](#): > > try it on the example. ```prolog ?- copy_to_singletons(foo(A, A), X). X = foo(_8998, _9000). ?- copy_to_singletons(foo(A,bar([A,zot(A)])), Z). Z = foo(_7662, bar([_7696, zot(_7732)])). ``` ---
**Author:** ![ridgeworks](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/ridgeworks/32/886_2.png) [@ridgeworks](https://swi-prolog.discourse.group/u/ridgeworks)\ **Post date:** [July 27, 2021, 4:48pm UTC](https://swi-prolog.discourse.group/t/copy-term-but-without-maintaining-variable-relationships/4204/11 "2021-07-27T16:48:48Z")
> [@peter.ludemann](#): > > ```prolog > ?- copy_to_singletons(foo(A, A), X). > X = foo(_8998, _9000). > > ``` That’s odd because my version doesn’t terminate: ```prolog ?- copy_to_singletons(foo(A,A),X). Action (h for help) ? goals [1,572,043] copy_to_singletons([_7854, [[[[[[[...]]]]]]]], _31448902) [1,572,042] copy_to_singletons([_7854, [[[[[[[...]]]]]]]], _31448882) [1,572,041] copy_to_singletons([_7854, [[[[[[[...]]]]]]]], _31448862) [1,572,040] copy_to_singletons([_7854, [[[[[[[...]]]]]]]], _31448842) [1,572,039] copy_to_singletons([_7854, [[[[[[[...]]]]]]]], _31448822) Action (h for help) ? abort % Execution Aborted ``` This is because the argument list of the compound term `[A,A]` is `[A,[A]]`, i.e. indefinite. ```prolog ?- [A,A] =.. L. L = ['[|]', A, [A]]. ``` So there’s something going on here that I’m not understanding. ---
**Author:** ![EricGT](https://avatars.discourse-cdn.com/v4/letter/e/f1d935/32.png) [@EricGT](https://swi-prolog.discourse.group/u/EricGT)\ **Post date:** [July 28, 2021, 1:53pm UTC](https://swi-prolog.discourse.group/t/copy-term-but-without-maintaining-variable-relationships/4204/12 "2021-07-28T13:53:11Z")
> [@ridgeworks](#): > > Am I missing something? Bumped into [explain.pl](https://github.com/SWI-Prolog/swipl-devel/blob/9cf6502af6fbac5a078dd670f07e1b13905c1263/library/explain.pl) and thought of your question. HTH