在下面的示例代码中,我想从[[:a :A] [:b :B] [:c :C]]
中删除[:b :B]
,如果用(lvar)
替换:B
,它将不再工作:
;; Helper Function
(defne not-membero [x l]
([_ []])
([_ [?y . ?r]]
(!= x ?y)
(not-membero x ?r)))
这些工作:
(run* [q]
(membero q [[:a :A] [:b :B] [:c :C]])
(not-membero q [[:b :B]]))
(run* [q]
(membero q [[:a :A] [:b :B] [:c :C]])
(!= q [:b :B]))
;; both return [[:a :A] [:c :C]], as expected
这些没有(注意lvar
):
(run* [q]
(membero q [[:a :A] [:b :B] [:c :C]])
(not-membero q [[:b (lvar)]]))
(run* [q]
(membero q [[:a :A] [:b :B] [:c :C]])
(!= q [:b (lvar)]))
;; both return [[:a :A] [:b :B] [:c :C]], unexpected
我认为这在您的示例中不起作用的原因是创建的(lvar)
与程序中的任何其他逻辑变量都是未绑定/无关的。如果您使用fresh
逻辑变量,您的程序运行良好(至少我认为这是您想要的)
(run* [q]
(fresh [x]
(membero q [[:a :A] [:b x] [:c :C]])
(not-membero q [[:b x]])))
=> ([:a :A] [:c :C])
(run* [q]
(fresh [x]
(membero q [[:a :A] [:b x] [:c :C]])
(!= q [:b x])))
=> ([:a :A] [:c :C])
或者,在不知道元组中的:b
项的情况下,这些返回相同的结果:
(run* [q]
(fresh [x]
(membero q [[:a :A] x [:c :C]])
(not-membero q [x])))
(run* [q]
(fresh [x]
(membero q [[:a :A] x [:c :C]])
(!= q x)))