All Classes and Interfaces
Class
Description
Base class for array states that track which group ids have been set.
Abstract base class for block types that are implemented by delegating to several concrete sub blocks.
Base class that extends
AbstractArrayState to add failure tracking.Releasable, non-threadsafe version of
AbstractRefCounted.Abstract superclass for operators that accept a single page, modify it, and then return it.
Maps a single
Page into zero or more resulting pages.Block that stores aggregate_metric_double values.
Literal to represent AggregateMetricDouble and primarily used for testing and during folding.
Blocking aggregation operator.
A non-grouped aggregation.
Builds aggregation implementations, closing over any state required to do so.
"Modes" for running an aggregate function.
A time-series aggregation function that collects the First occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the First occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the First occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the First occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the First occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the First occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the First occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the First occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the First occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the First occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the First occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the First occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the Last occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the Last occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the Last occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the Last occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the Last occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the Last occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the Last occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the Last occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the Last occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the Last occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the Last occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the Last occurrence value of a time series in a specified interval.
An implementation which always returns
RefCounted.ALWAYS_REFERENCED for any shard ID.An aggregator that retrieves the first value it encounters.
An aggregator that retrieves the first value it encounters.
An aggregator that retrieves the first value it encounters.
An aggregator that retrieves the first value it encounters.
An aggregator that retrieves the first value it encounters.
An aggregator that retrieves the first value it encounters.
Constructor for a block backed by Arrow buffers.
Constructor for a vector backed by Arrow buffers.
RowInTableLookup that models an increasing sequence of integers.AsyncOperator performs an external computation specified in
AsyncOperator.performAsync(Page, ActionListener).Shared context for batch state management between BatchDriver and PageToBatchPageOperator.
Batch lifecycle states.
Driver that processes batches on the server side of a BidirectionalBatchExchange.
Decodes values encoded by
BatchEncoder.Checks if an offset is
null.Request sent from client to server to track batch exchange status.
Response sent from server to client indicating batch exchange completion status.
Metadata for batch processing in bidirectional batch exchanges.
Wraps an
ExchangeSource and reorders BatchPages within each batch
to ensure they are output in sequential order based on pageIndexInBatch.Client-side handler for bidirectional batch exchange.
Functional interface for server setup callback.
Server-side handler for bidirectional batch exchange.
A Block is a columnar representation of homogenous data.
Builds
Blocks.How are multivalued fields ordered?
Some operators can enable its optimization when mv_values are sorted ascending or de-duplicated.
Builder for
BlockFactory instances.Specialized hash table implementations that map rows to a set
of bucket IDs to which they belong to implement
GROUP BY expressions.Configuration for a BlockHash group spec that is doing text categorization.
Configuration for a BlockHash group spec that is later sorted and limited (Top-N).
Represents the optimization state for block processing in lookup/enrich operations.
Stream for reading
Blocks.Returned by
BlockUtils.toJavaObject(org.elasticsearch.compute.data.Block, int) for "doc" type blocks.Block implementation that stores values in a
BooleanArrayVector.Arrow buffer backed
BooleanBlock.Arrow buffer backed
BooleanVector.Block implementation that stores values in a
BooleanBigArrayVector.Vector implementation that defers to an enclosed
BitArray.Block that stores boolean values.
Builder for
BooleanBlockAggregates the top N boolean values per bucket.
Vector that stores boolean values.
A builder that grows as needed.
A builder that never grows.
Block view of a
BooleanVector.Builder for
BooleanVectors that never grows.Builder for bytes arrays that checks its size against a
CircuitBreaker.Can be used to deep copy or generate data for
TDigestHolders with correct memory accounting.Block implementation that stores values in a
BytesRefArrayVector.Aggregator state for an array of BytesRefs.
Arrow buffer backed
BytesRefBlock.Arrow buffer backed
BytesRefVector.Block that stores BytesRef values.
Builder for
BytesRefBlockAggregates the top N variable length
BytesRef values per bucket.Maps a
LongBlock column paired with a BytesRefBlock column to group ids.Vector that stores BytesRef values.
A builder that grows as needed.
Block view of a
BytesRefVector.BlockHash implementation for
Categorize grouping function.BlockHash implementation for
Categorize grouping function as first
grouping expression, followed by one or mode other grouping expressions.Find spikes, dips and change point in a list of values.
Arrow allocation listener that uses a circuit breaker to track memory usage.
A
BlockLoader.ColumnAtATimeReader that immediately closes the reader after it is used.Factory for
ColumnLoadOperator.A base class that loads in and works across all input pages for an operator
ExpressionEvaluators returning constants and their factories.Block implementation representing a constant null value.
This vector is never instantiated.
How we partition the data across
Drivers.A
TopNEncoder that doesn't encode values so they are sortable but is
capable of encoding any values.A rate grouping aggregation definition for double.
A rate grouping aggregation definition for float.
A rate grouping aggregation definition for int.
A rate grouping aggregation definition for long.
Specialized CountAggregatorFunction for dense_vectors.
Specialized
CountApproximateAggregatorFunction for dense_vectors.Specialized
CountApproximateGroupingAggregatorFunction for dense_vectors.Specialized CountGroupingAggregatorFunction for dense_vectors.
A component capable of describing itself.
Operator that filters rows to keep only the first occurrence of each distinct key value.
Vector where each entry references a lucene document.Block implementation that stores values in a
DoubleArrayVector.Arrow buffer backed DoubleBlock.
Arrow buffer backed DoubleVector.
Block implementation that stores values in a
DoubleBigArrayVector.Vector implementation that defers to an enclosed
DoubleArray.Block that stores double values.
Builder for
DoubleBlockAggregates the top N
double values per bucket.Aggregates the top N
double values per bucket.Aggregates the top N
double values per bucket.Aggregates the top N
double values per bucket.Aggregates the top N
double values per bucket.Vector that stores double values.
A builder that grows as needed.
A builder that never grows.
Block view of a
DoubleVector.Builder for
DoubleVectors that never grows.Information returned when one of more
Drivers is completed.A driver-local context that is shared across operators.
A snapshot of the driver context.
Indicates the behavior Evaluators of this context should use for reporting warnings
An exception indicates that a compute should be terminated early as the downstream pipeline has enough or no long requires more data.
Profile results from a single
Driver.Run a set of drivers to completion.
Records of the times the driver has slept.
A record of a time the driver slept.
Task.Status reported from a Driver to be reported by the tasks api.A
DriverRunner that executes Driver with a child task so that we can retrieve the progress with the Task API.RowInTableLookup for an empty table.Lookup document IDs for the input queries.
Similar to
RefCountingListener,
but prefers non-task-cancelled exceptions over task-cancelled ones as they are more useful for diagnosing issues.Evaluates a tree of functions for every position in the block, resulting in a
new block which is appended to the page.
ExchangeService is responsible for exchanging pages between exchange sinks and sources on the same or different nodes.Sink for exchanging data
An
ExchangeSinkHandler receives pages and status from its ExchangeSinks, which are created using
ExchangeSinkHandler.createExchangeSink(Runnable)} method.Sink operator implementation that pushes data to an
ExchangeSinkSource for exchanging data
An
ExchangeSourceHandler asynchronously fetches pages and status from multiple RemoteSinks
and feeds them to its ExchangeSource, which are created using the ExchangeSourceHandler.createExchangeSource()) method.Source operator implementation that retrieves data from an
ExchangeSourceA block that holds
ExponentialHistogram values.Builder for
ExponentialHistogramBlockAbstraction to use for reading individual serialized via
ExponentialHistogramBlockBuilder.deserializeAndAppend(SerializedInput).Abstraction to use for writing individual values via
ExponentialHistogramBlock.serializeExponentialHistogram(int, SerializedOutput, BytesRef).Reusable storage to be passed to
ExponentialHistogramBlock.getExponentialHistogram(int, ExponentialHistogramScratch).A grouping state consisting of a single
long value with a ExponentialHistogram per group.A state consisting of a single
long value with a ExponentialHistogram.A Factory for creating ExpressionEvaluators.
FailureCollector is responsible for collecting exceptions that occur in the compute engine.A
AggregatorFunctionSupplier that wraps another, filtering which positions
are supplied to the aggregator.A time-series aggregation function that collects the First occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the First occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the first occurrence exponential histogram of a time series in a specified interval.
A time-series aggregation function that collects the First occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the First occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the First occurrence value of a time series in a specified interval.
Block implementation that stores values in a
FloatArrayVector.Arrow buffer backed FloatBlock.
Arrow buffer backed FloatVector.
Block implementation that stores values in a
FloatBigArrayVector.Vector implementation that defers to an enclosed
FloatArray.Block that stores float values.
Builder for
FloatBlockAggregates the top N
float values per bucket.Aggregates the top N
float values per bucket.Aggregates the top N
float values per bucket.Aggregates the top N
float values per bucket.Aggregates the top N
float values per bucket.Vector that stores float values.
A builder that grows as needed.
A builder that never grows.
Block view of a
FloatVector.Builder for
FloatVectors that never grows.Holds the options that are given to the FUSE command.
An operator that limits the number of rows emitted per group.
A top-N operator for grouped (SORT + LIMIT BY) queries.
Provides a context for evaluating a grouping aggregator.
Applies some grouping function like
min or avg to many values,
grouped into buckets.Consume group ids to cause the
GroupingAggregatorFunction
to group values at a particular position into a particular group.View into the agg that's prepared to emit results.
Encodes the values at a given position across multiple blocks into a single
BytesRef composite key.Builder for
HashAggregationOperator.A factory for constructing concrete Hash implementations.
Exponential histograms are composite data types.
Ceci n'est pas une List (though it may be backed by one).
An
IndexedByShardId used by only in tests, but placed here so it's visible for all test modules.A simple implementation when there's only a single value being used.
Block implementation that stores values in a
IntArrayVector.Arrow buffer backed IntBlock.
Arrow buffer backed IntVector.
Block implementation that stores values in a
IntBigArrayVector.Vector implementation that defers to an enclosed
IntArray.Block that stores int values.
Builder for
IntBlockAggregates the top N
int values per bucket.Aggregates the top N
int values per bucket.Intermediate aggregation state descriptor.
Aggregates the top N
int values per bucket.Aggregates the top N
int values per bucket.Aggregates the top N
int values per bucket.Vector that stores int values.
A builder that grows as needed.
A builder that never grows.
Block view of a
IntVector.Builder for
IntVectors that never grows.Aggregator state for an array of IPs.
Aggregates the top N IP values per bucket.
A rate grouping aggregation definition for double.
A rate grouping aggregation definition for float.
A rate grouping aggregation definition for int.
A rate grouping aggregation definition for long.
Is this
Operator blocked?A time-series aggregation function that collects the Last occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the Last occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the Last occurrence exponential histogram of a time series in a specified interval.
A time-series aggregation function that collects the Last occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the Last occurrence value of a time series in a specified interval.
A time-series aggregation function that collects the Last occurrence value of a time series in a specified interval.
A shared limiter used by multiple drivers to collect hits in parallel without exceeding the output limit.
Updates the score column in two stages:
1.
ExpressionEvaluator that loads from the Page directly.Requesting and returning memory from a
CircuitBreaker can be costly due to the involvement of read/write
on one or several atomic longs.A
CircuitBreakerService that only supports a single LocalCircuitBreaker.Block implementation that stores values in a
LongArrayVector.Arrow buffer backed LongBlock.
Arrow buffer backed LongVector.
Block implementation that stores values in a
LongBigArrayVector.Vector implementation that defers to an enclosed
LongArray.Block that stores long values.
Builder for
LongBlockAggregates the top N
long values per bucket.Aggregates the top N
long values per bucket.Aggregates the top N
long values per bucket.Aggregates the top N
long values per bucket.Aggregates the top N
long values per bucket.Block that stores long ranges.
Aggregates the top N collected values, and keeps them sorted.
Vector that stores long values.
A builder that grows as needed.
A builder that never grows.
Block view of a
LongVector.Builder for
LongVectors that never grows.An interface to generates queries for the lookup and enrich operators.
Intermediate operator that processes match field pages from ExchangeSourceOperator
and generates queries to lookup document IDs.
Source operator that incrementally counts the results in Lucene searches
Returns always one entry that mimics the Count aggregation internal state:
1.
Factory that generates an operator that finds the max value of a field using the
LuceneMinMaxOperator.Factory that generates an operator that finds the min value of a field using the
LuceneMinMaxOperator.Base class for evaluating a Lucene query at the compute engine and providing a Block as a result.
ExpressionEvaluator to run a Lucene Query during
the compute engine's normal execution, yielding matches/does not match into
a BooleanVector.ScoreOperator.ExpressionScorer to run a Lucene Query during
the compute engine's normal execution, yielding the corresponding scores into
a DoubleVector.Holds a list of multiple partial Lucene segments
Shared Lucene slices between Lucene operators.
Strategy used to partition each shard into slices.
Query to run and tags to add to the results.
Source operator that incrementally runs Lucene searches
Source operator that builds Pages out of the output of a TopFieldCollector (aka TopN).
Combines values at the given blocks with the same positions into a single position
for the blocks at the given channels.
Operator for the ESQL
METRICS_INFO command.Factory for INITIAL mode (data nodes): extracts metric metadata from shards.
Factory for FINAL mode (coordinator): merges 6-column pages from multiple data nodes.
Looks up metric field metadata on demand.
ES|QL Operator for performing MMR result diversification
MMR performs result diversification on incoming results using maximum marginal relevance.
Factor creation for the MMR Operator
Utilities to remove duplicates from multivalued fields.
Result of calling "hash" on a multivalue dedupe.
Removes duplicate values from multivalued positions.
Removes duplicate values from multivalued positions.
Removes duplicate values from multivalued positions.
Removes duplicate values from multivalued positions.
Removes duplicate values from multivalued positions.
"Expands" multivalued blocks by duplicating all the other columns for each value.
Operator is low-level building block that consumes, transforms and produces data.
A factory for creating intermediate operators.
Status of an
Operator to be returned by the tasks API.Status of an
Operator.A
BytesRefBlock consists of a pair: an IntBlock for ordinals and a BytesRefVector for the dictionary.A
BytesRefVector consists of a pair: an IntVector for ordinals and a BytesRefVector for the dictionary.Sink operator that calls a given listener for each page received.
Interface for DriverContext implementations that provide a collectedPages list.
A page is a column-oriented data abstraction that allows data to be passed between operators in
batches.
Sink operator that's useful for passing off pages to a
Consumer.Wrapper around
CloseableIterator of Pages that stops yielding
pages once a cumulative row budget is exhausted.A subset of a
LeafReaderContext.Profile information for plan optimization phases.
Generates a list of Lucene queries based on the input block.
A helper class that can be used to collect and merge response headers from multiple child requests.
Performs a
LEFT JOIN where many "right hand" pages are joined
against a "left hand" Page.Consumes
Pages and looks up each row in a pre-built table, and returns the
offsets of each row in the table.Factory for
RowInTableLookupOperator.Updates the score column with new scores using the RRF formula.
Evaluates scores for a ExpressionScorer.
Evaluates the score of an expression one
Page at a time.A Factory for creating ExpressionScorers.
Context of each shard we're operating against.
A thread safe, shared holder for the min competitive value from a
set of
TopNOperators.A "side channel" for the
Operators running on one node.Builds the
SideChannel on first use and returns a shared reference
on the next use.Like
DoubleBlockBuilder but optimized for collecting dense single valued values.Like
IntBlockBuilder but optimized for collecting dense single valued values.Like
LongBlockBuilder but optimized for collecting dense single valued values.Finds all fields with a single-value.
A sink operator - accepts input, produces no output.
A factory for creating sink operators.
A
TopNEncoder that encodes values to byte arrays that may be sorted directly.A
TopNEncoder that encodes values to byte arrays that may be sorted directly.A source operator - produces output, accepts no input.
A factory for creating source operators.
This operator generates 0 values for all buckets that are not present in the input data for a sparkline graph.
A standard deviation aggregation definition for double.
A standard deviation aggregation definition for float.
A standard deviation aggregation definition for int.
A standard deviation aggregation definition for long.
Builder for
TDigestBlockThis is a
TDigestReadView annotated with some extra information: The sum, min and max of all observations.A grouping state consisting of a single
long value with a TDigestHolder per group.A state consisting of a single
long value with a TDigestHolder.A specialized version of
HashAggregationOperator that aggregates time-series aggregations from time-series sources.An optimized block hash that receives two blocks: tsid and timestamp, which are sorted.
Extension of
LuceneSourceOperator for time-series aggregation that inserts metadata blocks,
such as slice index and future max timestamp, to allow downstream operators to optimize processing.Result from calling
BooleanBlock.toMask().An internal aggregate function that always emits intermediate (or partial) output regardless of the aggregate mode.
Encodes values for
TopNOperator.Removes duplicate values from multivalued positions, and keeps only the ones that pass the filters.
An operator that sorts "rows" of values by encoding the values to sort on, as bytes (using BytesRef).
Operator for the ESQL
TS_INFO command.Factory for INITIAL mode (data nodes).
Factory for FINAL mode (coordinator): merges 7-column pages from multiple data nodes.
Helper for the
ValuesIntAggregator, ValuesFloatAggregator and ValuesBytesRefAggregator
collecting values after the first in each group are collected in a hash, keyed
by the pair of groupId and value.Helper for the
ValuesLongAggregator and ValuesDoubleAggregator
collecting values after the first in each group are collected in a hash, keyed
by the pair of groupId and value.Operator that extracts doc_values from a Lucene index out of pages that have been produced by
LuceneSourceOperator
and outputs them to a new column.Builds a
ValuesSourceReaderOperator.LoaderAndConverter for a given shard.Evaluator for any type conversions that must be performed on load.
Creates a factory for
ValuesSourceReaderOperator.Configuration for a field to load.
A dense Vector of single values.
Builds
Vectors.Utilities to collect warnings for running an executor.
Interface for providing source location information for warnings.
Algorithm for calculating standard deviation, one value at a time.
A
AggregatorFunctionSupplier that wraps another, and apply a window function on the final aggregation.A
GroupingAggregatorFunction that wraps another, and apply a window function on the final aggregation.