Find only numbers in list

问题:

I need to write a rule, what finds only numbers in list M and outputs only the numbers in list O.

Query looks like: find(M, O)

I can't figure it out on my own and hope someone can help me.

回答1:

Here is a "classical Prolog" way to do it:

find([], []).
find([H|T], [H|NewT]) :-
    number(H),
    find(T, NewT).
find([H|T], NewT) :-
    \+ number(H),
    find(T, NewT).

There are 3 clauses.

1st says that for an empty list the result is an empty list.

2nd says: if the first element (head, H) of the input list is a number, keep it in the output, and the rest of the output (new tail, NewT) is find applied to the rest of the input list (tail, T).

3rd clause is structurally similar to the second, but says not to keep head if it's not a number.

回答2:

Here's another approach, using findall/3 which is available in most Prolog implementations. But for a beginner, Sergey's answer is arguably best for beginners to learn how basic recursive list processing works.

find(M, O) :-
    findall(X, (member(X, M), number(X)), O).

回答3:

If you're using swi-prolog, you'll be happy to use the include/3 predicate as so:

find_numbers(M,O) :-
    include(number,M,O).