Struct ScanInput

pub(crate) struct ScanInput {

    access_layer: AccessLayerRef,
    pub(crate) mapper: Arc<ProjectionMapper>,
    time_range: Option<TimestampRange>,
    pub(crate) predicate: PredicateGroup,
    pub(crate) memtables: Vec<MemRangeBuilder>,
    pub(crate) files: Vec<FileHandle>,
    pub(crate) cache_strategy: CacheStrategy,
    ignore_file_not_found: bool,
    pub(crate) parallel_scan_channel_size: usize,
    inverted_index_applier: Option<Arc<InvertedIndexApplier>>,
    bloom_filter_index_applier: Option<Arc<BloomFilterIndexApplier>>,
    fulltext_index_applier: Option<Arc<FulltextIndexApplier>>,
    pub(crate) query_start: Option<Instant>,
    pub(crate) append_mode: bool,
    pub(crate) filter_deleted: bool,
    pub(crate) merge_mode: MergeMode,
    pub(crate) series_row_selector: Option<TimeSeriesRowSelector>,
    pub(crate) distribution: Option<TimeSeriesDistribution>,
}

Common input for different scanners.

Fields

access_layer: AccessLayerRef

Region SST access layer.

mapper: Arc<ProjectionMapper>

Maps projected Batches to RecordBatches.

time_range: Option<TimestampRange>

Time range filter for time index.

predicate: PredicateGroup

Predicate to push down.

memtables: Vec<MemRangeBuilder>

Memtable range builders for memtables in the time range..

files: Vec<FileHandle>

Handles to SST files to scan.

cache_strategy: CacheStrategy

Cache.

ignore_file_not_found: bool

Ignores file not found error.

parallel_scan_channel_size: usize

Capacity of the channel to send data from parallel scan tasks to the main task.

inverted_index_applier: Option<Arc<InvertedIndexApplier>>

Index appliers.

bloom_filter_index_applier: Option<Arc<BloomFilterIndexApplier>>

fulltext_index_applier: Option<Arc<FulltextIndexApplier>>

query_start: Option<Instant>

Start time of the query.

append_mode: bool

The region is using append mode.

filter_deleted: bool

Whether to remove deletion markers.

merge_mode: MergeMode

Mode to merge duplicate rows.

series_row_selector: Option<TimeSeriesRowSelector>

Hint to select rows from time series.

distribution: Option<TimeSeriesDistribution>

Hint for the required distribution of the scanner.

Implementations

impl ScanInput

pub(crate) fn new( access_layer: AccessLayerRef, mapper: ProjectionMapper, ) -> ScanInput

Creates a new ScanInput.

pub(crate) fn with_time_range(self, time_range: Option<TimestampRange>) -> Self

Sets time range filter for time index.

pub(crate) fn with_predicate(self, predicate: PredicateGroup) -> Self

Sets predicate to push down.

pub(crate) fn with_memtables(self, memtables: Vec<MemRangeBuilder>) -> Self

Sets memtable range builders.

pub(crate) fn with_files(self, files: Vec<FileHandle>) -> Self

Sets files to read.

pub(crate) fn with_cache(self, cache: CacheStrategy) -> Self

Sets cache for this query.

pub(crate) fn with_ignore_file_not_found(self, ignore: bool) -> Self

Ignores file not found error.

pub(crate) fn with_parallel_scan_channel_size( self, parallel_scan_channel_size: usize, ) -> Self

Sets scan task channel size.

pub(crate) fn with_inverted_index_applier( self, applier: Option<Arc<InvertedIndexApplier>>, ) -> Self

Sets invereted index applier.

pub(crate) fn with_bloom_filter_index_applier( self, applier: Option<Arc<BloomFilterIndexApplier>>, ) -> Self

Sets bloom filter applier.

pub(crate) fn with_fulltext_index_applier( self, applier: Option<Arc<FulltextIndexApplier>>, ) -> Self

Sets fulltext index applier.

pub(crate) fn with_start_time(self, now: Option<Instant>) -> Self

Sets start time of the query.

pub(crate) fn with_append_mode(self, is_append_mode: bool) -> Self

pub(crate) fn with_filter_deleted(self, filter_deleted: bool) -> Self

Sets whether to remove deletion markers during scan.

pub(crate) fn with_merge_mode(self, merge_mode: MergeMode) -> Self

Sets the merge mode.

pub(crate) fn with_distribution( self, distribution: Option<TimeSeriesDistribution>, ) -> Self

Sets the distribution hint.

pub(crate) fn with_series_row_selector( self, series_row_selector: Option<TimeSeriesRowSelector>, ) -> Self

Sets the time series row selector.

pub(crate) fn create_parallel_sources( &self, sources: Vec<Source>, semaphore: Arc<Semaphore>, ) -> Result<Vec<Source>>

Scans sources in parallel.

Panics if the input doesn’t allow parallel scan.

pub(crate) fn build_mem_ranges( &self, index: RowGroupIndex, ) -> SmallVec<[MemtableRange; 2]>

Builds memtable ranges to scan by index.

pub(crate) async fn prune_file( &self, file_index: usize, reader_metrics: &mut ReaderMetrics, ) -> Result<FileRangeBuilder>

Prunes a file to scan and returns the builder to build readers.

pub(crate) fn spawn_scan_task( &self, input: Source, semaphore: Arc<Semaphore>, sender: Sender<Result<Batch>>, )

Scans the input source in another task and sends batches to the sender.

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

pub(crate) fn predicate(&self) -> Option<&Predicate>

Returns table predicate of all exprs.

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

Returns number of memtables to scan.

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

Returns number of SST files to scan.