Haskell#splitAt

Question

Here's how I wrote the Haskell splitAt function:

splitAt' :: Int -> [a] -> ([a], [a])
splitAt' n ys 
 | n < 0 = ([], ys)
 | otherwise = splitAt'' n ys []
 where splitAt'' a (x:xs) acc
 | a == 0 = (acc, x:xs)
 | null xs = (acc ++ [x], [])
 | otherwise = splitAt'' (a-1) xs (acc ++ [x])

I don't like I'm using the append (++) function to add an element to the end of my acc(umulator).

But, given the importance of ordering, I'm not sure how to avoid using it.

Please review this code as well.

Answer 1

You should be able to implement it without ++.

splitAt' :: Int -> [a] -> ([a], [a])
splitAt' 0 ys = ([], ys)
splitAt' _ [] = ([], [])
splitAt' n (y:ys)
 | n < 0 = ([], (y:ys))
 | otherwise = ((y:a), b)
 where (a, b) = splitAt' (n - 1) ys

• Kevin Meredith: Hi @200_success. Re: negatives aren't allowed: *Main> splitAt (-3) "heyman" produces ("","heyman") using GHCI 7.6.3. Also, with your above function, using a negative won't produce an error. It'll simply put the entire list in the first part of the tuple, leaving [] for the second part. Example: *Main> splitAt' (-3) [1..100] produces ([1,...,100],[])
• 200_success: You're right! I've added a case to handle negative n.
• Kevin Meredith: Thanks! Out of curiosity, when do you use the "don't care" (_) for arguments? Example: splitAt' n [] but n isn't used so why not use don't care? Maybe there's a general style rule?
• 200_success: Right again! The "don't care" placeholder should be used whenever possible. I've edited that in.