Safe Haskell | None |
---|---|
Language | Haskell2010 |
Synopsis
- newtype ParquetSource m = ParquetSource (Integer -> ConduitT () ByteString m ())
- type Url = String
- readMetadata :: (MonadError Text m, MonadIO m) => ParquetSource m -> m FileMetadata
- localParquetFile :: MonadResource m => FilePath -> ParquetSource m
- remoteParquetFile :: (MonadResource m, MonadThrow m, MonadIO m) => Url -> ParquetSource m
- readWholeParquetFile :: (MonadThrow m, MonadIO m, MonadError Text m, MonadResource m, MonadLogger m) => String -> m [ParquetObject]
- type Record = [(ColumnValue, [Text])]
- sourceParquet :: FilePath -> ConduitT () ParquetObject (ResourceT IO) ()
- sourceRowGroupFromRemoteFile :: (MonadResource m, MonadIO m, MonadThrow m, MonadLogger m) => String -> FileMetadata -> RowGroup -> ConduitT () ParquetObject m ()
- sourceRowGroup :: forall m. (MonadResource m, MonadIO m, MonadThrow m, MonadLogger m) => ParquetSource m -> FileMetadata -> RowGroup -> ConduitT () ParquetObject m ()
- sourceColumnChunk :: (MonadIO m, MonadResource m, MonadThrow m, MonadLogger m) => ParquetSource m -> FileMetadata -> ColumnChunk -> ConduitT () ColumnValue m ()
Documentation
newtype ParquetSource m Source #
ParquetSource (Integer -> ConduitT () ByteString m ()) |
readMetadata :: (MonadError Text m, MonadIO m) => ParquetSource m -> m FileMetadata Source #
localParquetFile :: MonadResource m => FilePath -> ParquetSource m Source #
remoteParquetFile :: (MonadResource m, MonadThrow m, MonadIO m) => Url -> ParquetSource m Source #
readWholeParquetFile :: (MonadThrow m, MonadIO m, MonadError Text m, MonadResource m, MonadLogger m) => String -> m [ParquetObject] Source #
type Record = [(ColumnValue, [Text])] Source #
sourceParquet :: FilePath -> ConduitT () ParquetObject (ResourceT IO) () Source #
sourceRowGroupFromRemoteFile :: (MonadResource m, MonadIO m, MonadThrow m, MonadLogger m) => String -> FileMetadata -> RowGroup -> ConduitT () ParquetObject m () Source #
sourceRowGroup :: forall m. (MonadResource m, MonadIO m, MonadThrow m, MonadLogger m) => ParquetSource m -> FileMetadata -> RowGroup -> ConduitT () ParquetObject m () Source #
Streams the values for every column chunk and zips them into records.
Illustration:
_____________________ | col1 | col2 | col3 | | 1 | a | x | | 2 | b | y | | 3 | c | z | |____|__|____|
sourceRowGroup
yields the following values in a stream:
(1, a, x) (2, b, y) (3, c, z)
sourceColumnChunk :: (MonadIO m, MonadResource m, MonadThrow m, MonadLogger m) => ParquetSource m -> FileMetadata -> ColumnChunk -> ConduitT () ColumnValue m () Source #