How can I run multiple Haskell IO actions and store the results?

Question

I'm a Haskell beginner, so sorry in advance for the newbie question. I only have a very superficial understanding of monads.

I'm using the function insert from the module Persistent. (I've been following the tutorial here .) It inserts something in a database, and returns the ID. I can use it like this:

resultId <- insert myItem

That works fine for a single item. I can print out the resultId like this:

liftIO $ print resultId

But what if my myItem is actually a list of arbitrary length? I want to map insert over this list, which I can seem to do with:

resultIds <- mapM_ insert myItemList

but then if I try to print out the values:

liftIO $ print resultIds

I just get () . What am I doing wrong?

Answer 1

You are quite close: you need mapM :: Monad m => (a -> mb) -> ta -> m (tb) instead of mapM_ :: (Foldable t, Monad m) => (a -> mb) -> ta -> m ()

Like the name and the signature already suggest, both functions take a monadic function and a traversable (let us for now assume that that is a list) of a s, and it applies the monadic function to all elements and returns a monadic function that contains a traversable (list) of the results.

So if you write:

    resultIds <-  insert myItemList

The difference between mapM and mapM_ is that in the case of mapM_ (like the signature already suggests), you are not interested in the outcome, and thus it is not calculated. A reason for this could possibly be that the list is very long (and generated by-need ), and thus the list of identifiers would never fit in memory.

then resultIds will contain a list of identifiers.

The explanation about the mapM (and mapM_ ) function is a bit an oversimplification, but I think that it is usually better to first get more comfortable about monads, than getting details about monadic functions completely right.