Safe Haskell	None
Language	Haskell2010

DataFrame.Lazy.IO.CSV

Synopsis

data ReadOptions = ReadOptions {
- hasHeader :: Bool
- inferTypes :: Bool
- safeRead :: Bool
- 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 :: 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 :: Bool rowRange :: !(Maybe (Int, Int)) seekPos :: !(Maybe Integer) totalRows :: !(Maybe Int) leftOver :: !Text rowsRead :: !Int

defaultOptions :: ReadOptions Source #

By default we assume the file has a header, we infer the types on read and we convert any rows with nullish objects into Maybe (safeRead).

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 :: 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 CSV/separated file for streaming, returning an open handle (positioned just after the header line) and the column specification for the schema columns that appear in the file header.

The caller is responsible for closing the handle when done.

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

Read up to batchSz rows from the open handle, returning a batch DataFrame and the unconsumed leftover text. Returns Nothing when the handle is at EOF and there is no leftover input.

The caller must pass the leftover returned by the previous call (use "" for 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.