Difference between revisions of "Floyd's cycle-finding algorithm"

From HaskellWiki
Jump to navigation Jump to search
(elements must be unique, e.g. findCycle $ cycle [0,1,0])
m (Fix typo)
 
Line 1: Line 1:
This is a Haskell impelementation of [http://en.wikipedia.org/wiki/Floyd's_cycle-finding_algorithm Floyd's cycle-finding algorithm] for finding cycles in lists.
+
This is a Haskell implementation of [http://en.wikipedia.org/wiki/Floyd's_cycle-finding_algorithm Floyd's cycle-finding algorithm] for finding cycles in lists.
 
<haskell>
 
<haskell>
 
findCycle :: Eq a => [a] -> ([a],[a])
 
findCycle :: Eq a => [a] -> ([a],[a])

Latest revision as of 00:39, 5 February 2019

This is a Haskell implementation of Floyd's cycle-finding algorithm for finding cycles in lists.

findCycle :: Eq a => [a] -> ([a],[a])
findCycle xxs = fCycle xxs xxs
  where fCycle (x:xs) (_:y:ys)
         | x == y              = fStart xxs xs
         | otherwise           = fCycle xs ys
        fCycle _      _        = (xxs,[]) -- not cyclic
        fStart (x:xs) (y:ys)
         | x == y              = ([], x:fLength x xs)
         | otherwise           = let (as,bs) = fStart xs ys in (x:as,bs)
        fLength x (y:ys)
         | x == y              = []
         | otherwise           = y:fLength x ys

This function is essentially the inverse of cycle. I.e. if xs and ys don't have a common suffix, their elements are unique and both are finite, we have that

findCycle (xs ++ cycle ys) == (xs,ys)