Text.Comma

The comma package implements parsing and producing of the CSV format.

Build & install

There are two standard ways of obtaining the module:

• by cloning the git repository: git clone https://github.com/lovasko/comma
• by using the central Hackage server: cabal install comma

Dependencies

Main part, the library itself, depends on two packages:

• attoparsec
• text

In addition, the testing component of the project depends on the QuickCheck package.

API

The Text.Comma module exports only two functions: comma and uncomma, parsing and producing CSV respectively. The function prototypes are as follows:

-- | Parse a block of text into a CSV table.
comma :: T.Text                   -- ^ CSV text
      -> Either String [[T.Text]] -- ^ error | table

-- | Render a table of texts into a valid CSV output.
uncomma :: [[T.Text]] -- ^ table
        -> T.Text     -- ^ CSV text

Example

The following example loads any CSV file and prepends a column that contains row numbers:

{-# LANGUAGE OverloadedStrings #-}

import Text.Comma
import System.Exit
import qualified Data.Text as T
import qualified Data.Text.IO as T

-- | Prepend a number in front of each row.
number :: [[T.Text]] -> [[T.Text]]
number = zipWith (:) (map (T.pack . show) [0..])

main :: IO ()
main = T.readFile "table.csv" >>= (\file -> case comma file of
  Left  err -> T.putStrLn ("ERROR: " <> T.pack err)           >> exitFailure
  Right csv -> T.writeFile "table.csv" (uncomma $ number csv) >> exitSuccess)

Example usage:

$ cat table.csv
name,surname,language
Dennis,Ritchie,C
Larry,Wall,Perl
John,McCarthy,Lisp
$ ./number && cat table.csv
0,name,surname,language
1,Dennis,Ritchie,C
2,Larry,Wall,Perl
3,John,McCarthy,Lisp

Standards

Both producing and parsing of CSV files closely follows the RFC4180 document. The comma implementation has made the following decisions:

• record separator is always only \n ('\r` is not recognized as a newline symbol)
• it is not required for the last record to be trailed with a \n. This means that comma "hello\nworld" == comma "hello\nworld\n" == Right [["hello"], ["world"]]

License

The comma package is licensed under the terms of the 2-clause BSD license. In case you need any other license, feel free to contact the author.

Author

Daniel Lovasko [email protected]