Struct GenericColumnWriter

pub struct GenericColumnWriter<'a, E: ColumnValueEncoder> {

    descr: ColumnDescPtr,
    props: WriterPropertiesPtr,
    statistics_enabled: EnabledStatistics,
    page_writer: Box<dyn PageWriter + 'a>,
    codec: Compression,
    compressor: Option<Box<dyn Codec>>,
    encoder: E,
    page_metrics: PageMetrics,
    column_metrics: ColumnMetrics<E::T>,
    encodings: BTreeSet<Encoding>,
    def_levels_sink: Vec<i16>,
    rep_levels_sink: Vec<i16>,
    data_pages: VecDeque<CompressedPage>,
    column_index_builder: ColumnIndexBuilder,
    offset_index_builder: Option<OffsetIndexBuilder>,
    data_page_boundary_ascending: bool,
    data_page_boundary_descending: bool,
    last_non_null_data_page_min_max: Option<(E::T, E::T)>,
}

Generic column writer for a primitive column.

Fields

descr: ColumnDescPtr

props: WriterPropertiesPtr

statistics_enabled: EnabledStatistics

page_writer: Box<dyn PageWriter + 'a>

codec: Compression

compressor: Option<Box<dyn Codec>>

encoder: E

page_metrics: PageMetrics

column_metrics: ColumnMetrics<E::T>

encodings: BTreeSet<Encoding>

The order of encodings within the generated metadata does not impact its meaning, but we use a BTreeSet so that the output is deterministic

def_levels_sink: Vec<i16>

rep_levels_sink: Vec<i16>

data_pages: VecDeque<CompressedPage>

column_index_builder: ColumnIndexBuilder

offset_index_builder: Option<OffsetIndexBuilder>

data_page_boundary_ascending: bool

data_page_boundary_descending: bool

last_non_null_data_page_min_max: Option<(E::T, E::T)>

(min, max)

Implementations

impl<'a, E: ColumnValueEncoder> GenericColumnWriter<'a, E>

pub fn new( descr: ColumnDescPtr, props: WriterPropertiesPtr, page_writer: Box<dyn PageWriter + 'a>, ) -> Self

Returns a new instance of GenericColumnWriter.

pub(crate) fn write_batch_internal( &mut self, values: &E::Values, value_indices: Option<&[usize]>, def_levels: Option<&[i16]>, rep_levels: Option<&[i16]>, min: Option<&E::T>, max: Option<&E::T>, distinct_count: Option<u64>, ) -> Result<usize>

pub fn write_batch( &mut self, values: &E::Values, def_levels: Option<&[i16]>, rep_levels: Option<&[i16]>, ) -> Result<usize>

Writes batch of values, definition levels and repetition levels. Returns number of values processed (written).

If definition and repetition levels are provided, we write fully those levels and select how many values to write (this number will be returned), since number of actual written values may be smaller than provided values.

If only values are provided, then all values are written and the length of of the values buffer is returned.

Definition and/or repetition levels can be omitted, if values are non-nullable and/or non-repeated.

pub fn write_batch_with_statistics( &mut self, values: &E::Values, def_levels: Option<&[i16]>, rep_levels: Option<&[i16]>, min: Option<&E::T>, max: Option<&E::T>, distinct_count: Option<u64>, ) -> Result<usize>

Writer may optionally provide pre-calculated statistics for use when computing chunk-level statistics

NB: WriterProperties::statistics_enabled must be set to EnabledStatistics::Chunk for these statistics to take effect. If EnabledStatistics::None they will be ignored, and if EnabledStatistics::Page the chunk statistics will instead be computed from the computed page statistics

pub(crate) fn memory_size(&self) -> usize

Returns the estimated total memory usage.

Unlike Self::get_estimated_total_bytes this is an estimate of the current memory usage and not the final anticipated encoded size.

pub fn get_total_bytes_written(&self) -> u64

Returns total number of bytes written by this column writer so far. This value is also returned when column writer is closed.

Note: this value does not include any buffered data that has not yet been flushed to a page.

pub(crate) fn get_estimated_total_bytes(&self) -> u64

Returns the estimated total encoded bytes for this column writer.

Unlike Self::get_total_bytes_written this includes an estimate of any data that has not yet been flushed to a page, based on it’s anticipated encoded size.

pub fn get_total_rows_written(&self) -> u64

Returns total number of rows written by this column writer so far. This value is also returned when column writer is closed.

pub fn get_descriptor(&self) -> &ColumnDescPtr

Returns a reference to a ColumnDescPtr

pub fn close(self) -> Result<ColumnCloseResult>

Finalizes writes and closes the column writer. Returns total bytes written, total rows written and column chunk metadata.

fn write_mini_batch( &mut self, values: &E::Values, values_offset: usize, value_indices: Option<&[usize]>, num_levels: usize, def_levels: Option<&[i16]>, rep_levels: Option<&[i16]>, ) -> Result<usize>

Writes mini batch of values, definition and repetition levels. This allows fine-grained processing of values and maintaining a reasonable page size.

fn should_dict_fallback(&self) -> bool

Returns true if we need to fall back to non-dictionary encoding.

We can only fall back if dictionary encoder is set and we have exceeded dictionary size.

fn should_add_data_page(&self) -> bool

Returns true if there is enough data for a data page, false otherwise.

fn dict_fallback(&mut self) -> Result<()>

Performs dictionary fallback. Prepares and writes dictionary and all data pages into page writer.

fn update_column_offset_index( &mut self, page_statistics: Option<&ValueStatistics<E::T>>, page_variable_length_bytes: Option<i64>, )

Update the column index and offset index when adding the data page

fn can_truncate_value(&self) -> bool

Determine if we should allow truncating min/max values for this column’s statistics

fn truncate_min_value( &self, truncation_length: Option<usize>, data: &[u8], ) -> (Vec<u8>, bool)

fn truncate_max_value( &self, truncation_length: Option<usize>, data: &[u8], ) -> (Vec<u8>, bool)

fn add_data_page(&mut self) -> Result<()>

Adds data page. Data page is either buffered in case of dictionary encoding or written directly.

fn flush_data_pages(&mut self) -> Result<()>

Finalises any outstanding data pages and flushes buffered data pages from dictionary encoding into underlying sink.

fn build_column_metadata(&mut self) -> Result<ColumnChunkMetaData>

Assembles column chunk metadata.

fn encode_levels_v1( &self, encoding: Encoding, levels: &[i16], max_level: i16, ) -> Vec<u8>

Encodes definition or repetition levels for Data Page v1.

fn encode_levels_v2(&self, levels: &[i16], max_level: i16) -> Vec<u8>

Encodes definition or repetition levels for Data Page v2. Encoding is always RLE.

fn write_data_page(&mut self, page: CompressedPage) -> Result<()>

Writes compressed data page into underlying sink and updates global metrics.

fn write_dictionary_page(&mut self) -> Result<()>

Writes dictionary page into underlying sink.

fn update_metrics_for_page(&mut self, page_spec: PageWriteSpec)

Updates column writer metrics with each page metadata.