Safe Haskell	None
Language	Haskell2010

DataFrame.Lazy.IO.CSV

Synopsis

data ReadOptions = ReadOptions {
- hasHeader :: Bool
- inferTypes :: Bool
- safeRead :: SafeReadMode
- safeReadOverrides :: [(Text, SafeReadMode)]
- rowRange :: !(Maybe (Int, Int))
- seekPos :: !(Maybe Integer)
- totalRows :: !(Maybe Int)
- leftOver :: !Text
- rowsRead :: !Int
}
defaultOptions :: ReadOptions
readCsv :: FilePath -> IO DataFrame
readTsv :: FilePath -> IO DataFrame
readSeparated :: Char -> ReadOptions -> FilePath -> IO (DataFrame, (Integer, Text, Int))
getInitialDataVectors :: Int -> IOVector MutableColumn -> [Text] -> IO ()
fillColumns :: Int -> Char -> IOVector MutableColumn -> IOVector [(Int, Text)] -> Text -> Handle -> IO (Text, Int)
writeValue :: IOVector MutableColumn -> IOVector [(Int, Text)] -> Int -> Int -> Text -> IO ()
freezeColumn :: Vector Text -> IOVector MutableColumn -> Vector [(Int, Text)] -> ReadOptions -> Int -> IO Column
openCsvStream :: Char -> Schema -> FilePath -> IO (Handle, [(Int, Text, SchemaType)])
readBatch :: Char -> [(Int, Text, SchemaType)] -> Int -> ByteString -> Handle -> IO (Maybe (DataFrame, ByteString))
writeColumnBs :: Int -> ByteString -> MutableColumn -> IO (Either Text Bool)
getNthFieldBs :: Word8 -> Int -> ByteString -> ByteString
makeCol :: Int -> SchemaType -> IO MutableColumn
sliceCol :: Int -> MutableColumn -> MutableColumn
findUnquotedNewline :: ByteString -> Maybe Int

Documentation

data ReadOptions Source #

Record for CSV read options.

Constructors

ReadOptions
Fields hasHeader :: Bool inferTypes :: Bool safeRead :: SafeReadMode Default `SafeReadMode` for columns without an entry in `safeReadOverrides`. safeReadOverrides :: [(Text, SafeReadMode)] Per-column `SafeReadMode` overrides; takes precedence over `safeRead`. rowRange :: !(Maybe (Int, Int)) seekPos :: !(Maybe Integer) totalRows :: !(Maybe Int) leftOver :: !Text rowsRead :: !Int

defaultOptions :: ReadOptions Source #

Default read options: assume a header, infer types, no safe-read wrapping. Set safeRead to MaybeRead/EitherRead to wrap columns; use safeReadOverrides to pick a different mode per column.

readCsv :: FilePath -> IO DataFrame Source #

Reads a CSV file from the given path. Note this file stores intermediate temporary files while converting the CSV from a row to a columnar format.

readTsv :: FilePath -> IO DataFrame Source #

Reads a tab separated file from the given path. Note this file stores intermediate temporary files while converting the CSV from a row to a columnar format.

readSeparated :: Char -> ReadOptions -> FilePath -> IO (DataFrame, (Integer, Text, Int)) Source #

Reads a character separated file into a dataframe using mutable vectors.

getInitialDataVectors :: Int -> IOVector MutableColumn -> [Text] -> IO () Source #

fillColumns :: Int -> Char -> IOVector MutableColumn -> IOVector [(Int, Text)] -> Text -> Handle -> IO (Text, Int) Source #

Reads rows from the handle and stores values in mutable vectors.

writeValue :: IOVector MutableColumn -> IOVector [(Int, Text)] -> Int -> Int -> Text -> IO () Source #

Writes a value into the appropriate column, resizing the vector if necessary.

freezeColumn :: Vector Text -> IOVector MutableColumn -> Vector [(Int, Text)] -> ReadOptions -> Int -> IO Column Source #

Freezes a mutable vector into an immutable one, trimming it to the actual row count.

openCsvStream :: Char -> Schema -> FilePath -> IO (Handle, [(Int, Text, SchemaType)]) Source #

Open a file for streaming: returns a handle positioned after the header and the column spec for the schema columns present in the header. The caller must close the handle when done.

readBatch :: Char -> [(Int, Text, SchemaType)] -> Int -> ByteString -> Handle -> IO (Maybe (DataFrame, ByteString)) Source #

Read up to batchSz rows, returning a batch DataFrame and the unconsumed leftover text; Nothing at EOF with no leftover. Pass the previous call's leftover back in (use "" on the first call).

writeColumnBs :: Int -> ByteString -> MutableColumn -> IO (Either Text Bool) Source #

Write a ByteString field value directly into a mutable column, parsing numerics without an intermediate Text allocation.

getNthFieldBs :: Word8 -> Int -> ByteString -> ByteString Source #

Extracts the Nth field (0-indexed), respecting double quotes and stripping them. Fast path: uses memchr-based break when no quotes are present in the line. Slow path: quote-aware character-by-character scan.

makeCol :: Int -> SchemaType -> IO MutableColumn Source #

Allocate a fresh MutableColumn for n slots based on a SchemaType.

sliceCol :: Int -> MutableColumn -> MutableColumn Source #

Slice a MutableColumn to n elements (no-copy view).

findUnquotedNewline :: ByteString -> Maybe Int Source #

Finds the index of the next unquoted newline (0x0A). Fast path: uses memchr (SIMD) and falls back to a quote-aware linear scan only if a double-quote appears before the candidate newline.