Recursion in a monad

Revision as of 21:47, 25 March 2009 by Lemming (Talk | contribs)

(diff) ←Older revision | Current revision (diff) | Newer revision→ (diff)

People sometimes wonder how to effectively do recursion when inside a

monadic

do

-block. Here's some quick examples:

The problem is to read 'n' lines from stdin, recursively:

The obvious, recursive way:

main = f 3
 
f 0 = return []
f n = do v  <- getLine
         vs <- f (n-1)
         return $! v : vs

Runs:

$ runhaskell A.hs
    1
    2
    3
    ["1","2","3"]

f 0 acc = return (reverse acc)
f n acc = do
    v  <- getLine
    f (n-1) (v : acc)

Or abstract the recursion pattern into a fold:

f n = do
    s <- foldM fn [] [1..n]
    return (reverse s)
 
  where fn acc _ = do x <- getLine
                      return (x:acc)

And finally, apply some functor and pointfree shortcuts:

f n = reverse `fmap` foldM fn [] [1..n]
    where fn acc _ = (: acc) `fmap` getLine