Question

I have two functions:

trans_table :: [Char] -> [Char] -> Map.Map Char Char
trans_table s1 s2 = Map.fromList (zip s1 s2)

random_trans_table :: IO (Map.Map Char Char)
random_trans_table = do
    rawKey <- shuffle ascii
    let zipped = zip ascii rawKey 
    let map = Map.fromList zipped
    return map    

First of them creates Map from two strings; second one generates random Map.
First one returns Map.Map Char Char; second one returns IO (Map.Map Char Char)

Now I need to lookup value from this Map and I have created two functions – one for IO Map and one for Map:

translate_char :: Map.Map Char Char -> Char -> Maybe Char
translate_char table c = Map.lookup c table

translate_char_io :: IO (Map.Map Char Char) -> Char -> IO (Maybe Char)
translate_char_io table c = do
    raw_table <- table
    let result = Map.lookup c raw_table
    return result

I don’t like it because it causes code duplication. I have already duplicate one function and if I code by this way I will need to duplicate all my functions.

Is there a way to create function which will works with Map and IO (Map) the same way ?

Answer 1

do-notation, or the “bind” operator >>= that it de-sugars to, takes care of this for you.

Anywhere that you would have called translate_char_io, you can use a call to the pure function by letting the monad syntax unwrap your table for you. For example, if you want to create a random table and look up two different characters in it, you could do it this way:

test c1 c2 = do
    table <- random_trans_table
    return (translate_char table c1, translate_char table c2)