Hi Frederic,
after investigating, I could not switch the IO and the Maybe I need to process the IO in order to know if I have a Just or a Nothing So this is a IO (Maybe ...)
I just would like to know if there is a better way to write this knowing
There's the MaybeT[1] monad transformer for this use case. Instead of having: row :: t -> Int -> IO (Maybe (DifTomoFrame DIM1)) you would have: row :: t -> Int -> MaybeT IO (DifTomoFrame DIM1) So 'row' might look like: row d idx = do n <- len d let eof = n - 1 == idx let nxs' = h5nxs d let mu = 0.0 let komega = 0.0 let kappa = 0.0 let kphi = 0.0 gamma <- get_position' (h5gamma d) 0 delta <- get_position' (h5delta d) idx wavelength <- get_position' (h5wavelength d) 0 let source = Source (head wavelength *~ nano meter) let positions = concat [mu, komega, kappa, kphi, gamma, delta] -- print positions let geometry = Geometry K6c source positions Nothing let detector = ZeroD m <- lift $ geometryDetectorRotationGet geometry detector poniext <- lift $ ponigen d (MyMatrix HklB m) idx return $ DifTomoFrame { difTomoFrameNxs = nxs' , difTomoFrameIdx = idx , difTomoFrameEOF = eof , difTomoFrameGeometry = geometry , difTomoFramePoniExt = poniext } This assumes that the functions 'len', 'get_position' also return a 'MaybeT IO (...)'. Functions in the IO monad - like 'geometryDetectorRotationGet' and 'ponigen' - have to be "lifted" into the IO monad by 'lift'. To get at the 'Maybe' result of 'row' you're using the 'runMaybeT' function in the IO monad: main :: IO () main = do ... result <- runMaybeT (row t int) ... Greetings, Daniel [1] https://hackage.haskell.org/package/transformers-0.5.2.0/docs/Control-Monad-...