Why does my haskell program keep running out of memory ?

Question

Hello fellow coders. So i decided to rewrite some of my old scripts i had lying around in haskell just because i need the practice and i like the language. So here i am trying to filter a huge file (around 1.7 GB) , cut the lines of no interest and write the remaining stuff in another file.

I thought that haskell's lazy nature would be ideal for this but the code keeps running out of memory too soon. The previous versions (c# or Python) had a read line -> write line approach but i tried a different approach here. Should i just rewrite the code to mirror the previous version or am i missing something.

So this is the function in charge of the original file filtering:

getLines :: FilePath -> IO [[String]]
getLines path = do
    text<-readFile path
    let linii=lines text
    let tokens = map words linii
    let filtrate=[x|x<-tokens,length x>7,isTimeStamp (x!!0),isDiagFrame x]
    return filtrate

this one is in charge of writing one line at a time in the new file (altho i tried to use writeFile dirrectly and failed miserably :) :

writeLines ::Handle->[[String]]->IO ()
writeLines handle linii = do
                    let linie=concat $ intersperse " " (head  linii)
                    hPutStrLn handle linie
                    if length linii > 0     then
                                    writeLines handle  (tail linii)
                                        else
                                    print "Writing complete..."

and these 2 are the main function and another one in charge of geting the handle and passing it around :

writeTheFile :: FilePath->FilePath->IO ()
writeTheFile inf outf = do
handle<-openFile outf WriteMode
linii<-getLines inf
writeLines handle linii
print "Write Complete"


main = do
arg<-getArgs
if length arg/=2    then
    print "Use like this : trace_pars [In_File] [Out_File] !"
                    else 
    writeTheFile (arg!!0) (arg!!1)

Any advice would be greatly appreciated...thanks in advance

dflemstr · Accepted Answer · 2012-02-08 10:13:27Z

14

The problem here is in this line:

                    if length linii > 0     then

You are computing the length of your list of lines. This means that the whole list of lines has to be loaded for it to be counted. Which means that the whole file that you're reading needs to be loaded into memory. Not good!

The solution is to use if not . null $ linii then instead. The null function checks whether a list is empty (which only forces the first line of the list to be loaded), and not behaves like you'd expect.

If you would like a more idiomatic version of writeLines (Note the use of FilePath instead of Handle):

writeLines :: FilePath -> [[String]] -> IO ()
writeLines filename = writeFile filename . unlines . map unwords

This function is the same as:

writeLines filename lines =
  writeFile filename mergedFile
  where
    mergedFile = unlines mergedLines
    mergedLines = map unwords lines

unlines is the same as intercalate "\n", and unwords is the same as intercalate " ". intercalate x is the same as concat . intersperse x.

I think that this should be enough information for you to understand what's going on.

answered Feb 8, 2012 at 10:13

dflemstr

26.3k5 gold badges74 silver badges107 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

omu_negru Over a year ago

thanks man....the condition was indeed the problem.....will try to do it without recursion next

Collectives™ on Stack Overflow

Why does my haskell program keep running out of memory ?

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related