System.IO.Streams.Csv.Tutorial

A simple tutorial on using the cassava-streams library to glue together cassava and io-streams.

Note: if you're reading this on Hackage or in Haddock then you should switch to source view with the "Source" link at the top of this page or open this file in your favorite text editor.

Types representing to-do items and their state

data Item

A to-do item.

Constructors

Item

Fields

• title :: String

  Title.

• state :: TState

  State.

• time :: Maybe Double

  Seconds taken to complete.

Instances

Eq Item

Methods

(==) :: Item -> Item -> Bool

(/=) :: Item -> Item -> Bool

Show Item

Methods

showsPrec :: Int -> Item -> ShowS

show :: Item -> String

showList :: [Item] -> ShowS

FromNamedRecord Item

Methods

parseNamedRecord :: NamedRecord -> Parser Item

ToNamedRecord Item

Methods

toNamedRecord :: Item -> NamedRecord

data TState

Possible states for a to-do item.

Constructors

Todo

Item needs to be completed.

Done

Item has been finished.

Instances

Eq TState

Methods

(==) :: TState -> TState -> Bool

(/=) :: TState -> TState -> Bool

Show TState

Methods

showsPrec :: Int -> TState -> ShowS

show :: TState -> String

showList :: [TState] -> ShowS

FromField TState

Methods

parseField :: Field -> Parser TState

ToField TState

Methods

toField :: TState -> Field

Functions which use cassava-streams functions

onlyTodo

Arguments

:: Handle

Input handle where CSV data can be read.

-> Handle

Output handle where CSV data can be written.

-> IO ()

The onlyTodo function reads to-do Items from the given input handle (in CSV format) and writes them back to the output handle (also in CSV format), but only if the items are in the Todo state. In another words, the CSV data is filtered so that the output handle only receives to-do Items which haven't been completed.

The io-streams handleToInputStream function is used to create an InputStream ByteString stream from the given input handle.

That stream is then given to the cassava-streams function decodeStreamByName which converts the InputStream ByteString stream into an InputStream Item stream.

Notice that the cassava-streams function onlyValidRecords is used to transform the decoding stream into one that only produces valid records. Any records which fail type conversion (via FromNamedRecord or FromRecord) will not escape from onlyValidRecords but instead will throw an exception.

Finally the io-streams filter function is used to filter the input stream so that it only produces to-do items which haven't been completed.

markDone

Arguments

:: String

Items with this title are marked as Done.

-> Handle

Input handle where CSV data can be read.

-> Handle

Output handle where CSV data can be written.

-> IO ()

The markDone function will read to-do items from the given input handle and mark any matching items as Done. All to-do items are written to the given output handle.