# Tabling: Problem with answer subsumption? **URL:** \ **Category:** Help!\ **Tags:** how-to\ **Created:** [July 11, 2020, 12:15am UTC](https://swi-prolog.discourse.group/t/tabling-problem-with-answer-subsumption/2624 "2020-07-11T00:15:43Z")\ **Posts on this page:** 10\ **Page:** 1
**Author:** ![swi](https://avatars.discourse-cdn.com/v4/letter/s/51bf81/32.png) [@swi](https://swi-prolog.discourse.group/u/swi)\ **Post date:** [July 11, 2020, 12:15am UTC](https://swi-prolog.discourse.group/t/tabling-problem-with-answer-subsumption/2624/1 "2020-07-11T00:15:43Z")
I have a small grammar that turns a number sequence, say `1,2,3,4,9` into a range/integer list: 1-4,9 Here is the code: ```prolog % :- table intrange_list(min,_,_). % Send bug report? :- table intrange_list//1. % Consolidate N, followed by (N+1)-H intrange_list([L-H]) --> integer(L), `,`, { L #= L1 - 1 }, intrange_list([L1-H]). % Basic N,N+1 sequence intrange_list([L-H]) --> integer(L), `,`, integer(H), { H #= L + 1 }. % Sequence followed by sequence or integer. intrange_list([L-H|Ls]) --> intrange_list([L-H]), `,`, intrange_list(Ls). % Integer followed by sequence or integer or by itself. intrange_list([i(N)|Ls]) --> integer(N), ( `,`, ! | [] ), intrange_list(Ls). % End of list. intrange_list([]) --> []. ``` Producing the following output: ```prolog 2 ?- phrase(intrange_list(L),`1,2,3,4,9`). L = [i(1), i(2), i(3), i(4), i(9)] ; L = [i(1), i(2), 3-4, i(9)] ; L = [i(1), 2-3, i(4), i(9)] ; L = [i(1), 2-4, i(9)] ; L = [1-3, i(4), i(9)] ; L = [1-2, i(3), i(4), i(9)] ; L = [1-2, 3-4, i(9)] ; L = [1-4, i(9)]. ``` I want to keep only the shortest answer (taking advantage of term ordering), so instead of using regular tabling I switched to: ```prolog :- table intrange_list(min,_,_). ``` But now I get this error: ```prolog 5 ?- phrase(intrange_list(L),`1,2,3,4,9`). ERROR: Uninstantiated argument expected, found [2-_16240] ERROR: In: ERROR: [21] throw(error(uninstantiation_error(...),_16290)) ... ``` Answer subsumption is not accepting the partially uninitialized term `[2-_]`, why not allow partially uninstantiated terms? Perhaps I am missing something. ---
**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 11, 2020, 7:51am UTC](https://swi-prolog.discourse.group/t/tabling-problem-with-answer-subsumption/2624/2 "2020-07-11T07:51:09Z")
> [@swi](#): > > Answer subsumption is not accepting the partially uninitialized term `[2-_]` , why not allow partially uninstantiated terms? That is not the problem. The clue is in **Un** _instantiated_ > [@swi](#): > > `ERROR: Uninstantiated argument expected, found [2-_16240]` The argument over which we do the answer subsumption must be unbound at call time. Anyway, it seems you want to minimize the list **length**. That is not going to work with standard order of term, which gives the list with the smallest first argument (and if equal the smallest second, etc.) ---
**Author:** ![swi](https://avatars.discourse-cdn.com/v4/letter/s/51bf81/32.png) [@swi](https://swi-prolog.discourse.group/u/swi)\ **Post date:** [July 11, 2020, 1:11pm UTC](https://swi-prolog.discourse.group/t/tabling-problem-with-answer-subsumption/2624/3 "2020-07-11T13:11:48Z")
Thanks, I ended up using `aggregate_all/3` with `set(..)` and took the shortest (last) solution. ---
**Author:** ![Boris](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/boris/32/7486_2.png) [@Boris](https://swi-prolog.discourse.group/u/Boris)\ **Post date:** [July 11, 2020, 2:00pm UTC](https://swi-prolog.discourse.group/t/tabling-problem-with-answer-subsumption/2624/4 "2020-07-11T14:00:35Z")
Just out of curiosity, why not use an algorithm that gives you the shortest solution first instead of last? You might even just take the first and be done? ---
**Author:** ![swi](https://avatars.discourse-cdn.com/v4/letter/s/51bf81/32.png) [@swi](https://swi-prolog.discourse.group/u/swi)\ **Post date:** [July 11, 2020, 7:06pm UTC](https://swi-prolog.discourse.group/t/tabling-problem-with-answer-subsumption/2624/5 "2020-07-11T19:06:14Z")
> [@Boris](#): > > st out of curiosity, why not use an algorithm that gives you the shortest solution first instead of last? I just put together the DCG in a few minutes, but you’re right, do you have an idea for a better DCG? ---
**Author:** ![Boris](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/boris/32/7486_2.png) [@Boris](https://swi-prolog.discourse.group/u/Boris)\ **Post date:** [July 11, 2020, 7:10pm UTC](https://swi-prolog.discourse.group/t/tabling-problem-with-answer-subsumption/2624/6 "2020-07-11T19:10:23Z")
It really depends what you are after. I thought this is just a minimal example that demonstrates a problem with tabling, and I don’t know about that so I assumed I am missing the big picture. So is the problem "transform the string ‘1,2,3,4,9’ into the list `[1-4,i(9)]`"? EDIT: and do you know that the original list of numbers is monotonically increasing? This makes it a bit easier I think. ---
**Author:** ![EricGT](https://avatars.discourse-cdn.com/v4/letter/e/f1d935/32.png) [@EricGT](https://swi-prolog.discourse.group/u/EricGT)\ **Post date:** [July 11, 2020, 7:12pm UTC](https://swi-prolog.discourse.group/t/tabling-problem-with-answer-subsumption/2624/7 "2020-07-11T19:12:16Z")
> [@swi](#): > > do you have an idea for a better DCG? This is not a DCG answer. Have you looked at [Run Length Encoding](https://en.wikipedia.org/wiki/Run-length_encoding) which is very similar. There is [working Prolog code for RLE](https://www.rosettacode.org/wiki/Run-length_encoding#Prolog) at RosettaCode. ---
**Author:** ![swi](https://avatars.discourse-cdn.com/v4/letter/s/51bf81/32.png) [@swi](https://swi-prolog.discourse.group/u/swi)\ **Post date:** [July 11, 2020, 7:15pm UTC](https://swi-prolog.discourse.group/t/tabling-problem-with-answer-subsumption/2624/8 "2020-07-11T19:15:56Z")
> [@Boris](#): > > So is the problem "transform the string ‘1,2,3,4,9’ into the list `[1-4,i(9)]` "? The problem is transform a sequence of integers into a string such that: 1. Any consecutive set of integers is shortened to Low-High 2. All the other integers are left in place EDIT: I need it to work both ways. This is very similar to what Eric pointed out: run length encoding. ---
**Author:** ![Boris](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/boris/32/7486_2.png) [@Boris](https://swi-prolog.discourse.group/u/Boris)\ **Post date:** [July 11, 2020, 7:38pm UTC](https://swi-prolog.discourse.group/t/tabling-problem-with-answer-subsumption/2624/9 "2020-07-11T19:38:48Z")
> [@swi](#): > > EDIT: I need it to work both ways. Too difficult for me, sorry. I never learned how to do that. ---
**Author:** ![swi](https://avatars.discourse-cdn.com/v4/letter/s/51bf81/32.png) [@swi](https://swi-prolog.discourse.group/u/swi)\ **Post date:** [July 11, 2020, 9:45pm UTC](https://swi-prolog.discourse.group/t/tabling-problem-with-answer-subsumption/2624/10 "2020-07-11T21:45:20Z")
Thanks Boris, you encourage me to put little bit more work, here is a version that works both ways (for the first answer) and gives the shortest answer first. ```prolog intlist_intranges([L-H|Ls]) --> { N1 #= L + 1 }, [L], lookahead(N1), intlist_intranges([N1-H|Ls]). intlist_intranges([L-H|Ls]) --> { H #= L + 1 }, [L], [H], intlist_intranges(Ls). intlist_intranges([i(I)|Ls]) --> [I], intlist_intranges(Ls). intlist_intranges([]) --> []. lookahead(A), [A] --> [A]. ```