how to parse yahoo historical csv with Attoparsec

Question

i am a beginner of haskell, how to parse with attoparsec into open array, high array etc

module CsvParser (
      Quote (..)
    , csvFile
    , quote
    ) where
import System.IO
import Data.Attoparsec.Text
import Data.Attoparsec.Combinator
import Data.Text (Text, unpack)
import Data.Time
import System.Locale
import Data.Maybe

data Quote = Quote {
        qTime       :: LocalTime,
        qAsk        :: Double,
        qBid        :: Double,
        qAskVolume  :: Double,
        qBidVolume  :: Double
    } deriving (Show, Eq)

csvFile :: Parser [Quote]
csvFile = do
    q <- many1 quote
    endOfInput
    return q

quote   :: Parser Quote
quote   = do
    time        <- qtime
    qcomma
    ask         <- double
    qcomma
    bid         <- double
    qcomma
    askVolume   <- double
    qcomma
    bidVolume   <- double
    endOfLine
    return $ Quote time ask bid askVolume bidVolume 

qcomma  :: Parser ()
qcomma  = do 
    char ','
    return ()

qtime   :: Parser LocalTime
qtime   = do
    tstring     <- takeTill (\x -> x == ',')
    let time    = parseTime defaultTimeLocale "%d.%m.%Y %H:%M:%S%Q" (unpack tstring)
    return $ fromMaybe (LocalTime (fromGregorian 0001 01 01) (TimeOfDay 00 00 00 )) time

--testString :: Text
--testString = "01.10.2012 00:00:00.741,1.28082,1.28077,1500000.00,1500000.00\n" 

quoteParser = parseOnly quote

main = do  
    handle <- openFile "C:\\Users\\ivan\\Downloads\\0005.HK.csv" ReadMode  
    contents <- hGetContents handle  
    let allLines = lines contents
    map (\line -> quoteParser line) allLines
    --putStr contents  
    hClose handle

Error message:

testhaskell.hs:89:5:
    Couldn't match type `[]' with `IO'
    Expected type: IO (Either String Quote)
      Actual type: [Either String Quote]
    In the return type of a call of `map'
    In a stmt of a 'do' block:
      map (\ line -> quoteParser line) allLines
    In the expression:
      do { handle <- openFile
                       "C:\\Users\\ivan\\Downloads\\0005.HK.csv" ReadMode;

           contents <- hGetContents handle;
           let allLines = lines contents;
           map (\ line -> quoteParser line) allLines;
           .... }

testhaskell.hs:89:37:
    Couldn't match type `[Char]' with `Text'
    Expected type: [Text]
      Actual type: [String]
    In the second argument of `map', namely `allLines'
    In a stmt of a 'do' block:
      map (\ line -> quoteParser line) allLines
    In the expression:
      do { handle <- openFile
                       "C:\\Users\\ivan\\Downloads\\0005.HK.csv" ReadMode;

           contents <- hGetContents handle;
           let allLines = lines contents;
           map (\ line -> quoteParser line) allLines;
           .... }

Answer 1

You can use attoparsec-csv package or you can take a look at its source code to have some idea on how to write it by yourself.

The code will be like

import qualified Data.Text.IO as T
import Text.ParseCSV

main = do
  txt <- T.readFile "file.csv"
  case parseCSV txt of
    Left  err -> error err
    Right csv -> mapM_ (print . mkQuote) csv

mkQuote :: [T.Text] -> Quote
mkQuote = error "Not implemented yet"

Answer 2

The error has nothing to do with parsec or attoparsec. The line the error message points to is not an IO action, so it causes the error when you try to use it as one:

main = do  
    handle <- openFile "C:\\Users\\ivan\\Downloads\\0005.HK.csv" ReadMode  
    contents <- hGetContents handle  
    let allLines = lines contents
    map (\line -> quoteParser line) allLines   -- <== This is not an IO action
    --putStr contents  
    hClose handl

You ignore the result of the map call. You should store it in a variable with let, like you do with the result of lines.

The second error is because you are trying to use Text as String which are different types, even though they both represent ordered collections of characters (they also have different internal representations). You can convert between the two types with pack and unpack: http://hackage.haskell.org/package/text/docs/Data-Text.html#g:5

Also, you should always explicitly give main the type signature main :: IO (). It can sometimes lead to subtle problems if you don't.

As other people have said, though, you should probably use a csv parser package.