Class SourceOperatorContext.Builder
java.lang.Object
org.elasticsearch.xpack.esql.datasources.spi.SourceOperatorContext.Builder
- Enclosing class:
SourceOperatorContext
-
Constructor Summary
Constructors -
Method Summary
Modifier and TypeMethodDescriptionattributes(List<Attribute> attributes) batchSize(int batchSize) build()datasetName(String datasetName) Registered dataset identifier (fromFROM <dataset>), ornullfor inlineEXTERNAL.declaredReadSpec(DeclaredReadSpec declaredReadSpec) The declared mapping's read-instructions (renames,_id.path), orDeclaredReadSpec.NONE.deferredExtraction(boolean deferredExtraction) Whether the plan pairs this source with anExternalFieldExtractExecconsuming deferred-encoded columns.fileReadExecutor(Executor fileReadExecutor) Optional executor for file (and similar) background reads.maxBufferSize(int maxBufferSize) maxConcurrentOpenSegments(int maxConcurrentOpenSegments) maxRecordBytes(int maxRecordBytes) parallelism(int parallelism) parsingParallelism(int parsingParallelism) partitionColumnNames(Set<String> partitionColumnNames) path(StoragePath path) projectedColumns(List<String> projectedColumns) pushedExpressions(List<Expression> pushedExpressions) pushedFilter(Object pushedFilter) rowLimit(int rowLimit) schemaMap(Map<StoragePath, SchemaReconciliation.FileSchemaInfo> schemaMap) Per-file planner-resolved schema info, populated by the resolver and threaded throughExternalSourceExec.sliceQueue(ExternalSliceQueue sliceQueue) sourceMetadata(Map<String, Object> sourceMetadata) sourceType(String sourceType) split(ExternalSplit split)
-
Constructor Details
-
Builder
public Builder()
-
-
Method Details
-
sourceType
-
path
-
projectedColumns
-
attributes
-
batchSize
-
maxBufferSize
-
rowLimit
-
executor
-
fileReadExecutor
Optional executor for file (and similar) background reads. When set (e.g. togeneric), async reads and slice-queue drain run here instead of onexecutor, so producers blocked in buffer backpressure do not starve theesql_workerdrivers that consume the buffer. -
config
-
sourceMetadata
-
pushedFilter
-
pushedExpressions
-
fileList
-
schemaMap
public SourceOperatorContext.Builder schemaMap(Map<StoragePath, SchemaReconciliation.FileSchemaInfo> schemaMap) Per-file planner-resolved schema info, populated by the resolver and threaded throughExternalSourceExec. Always present for resolved sources (single-file gets a one-entry identity map; multi-file gets the reconciliation result). Empty map for legacy/unresolved paths. -
split
-
partitionColumnNames
-
sliceQueue
-
parsingParallelism
-
maxConcurrentOpenSegments
-
maxRecordBytes
-
parallelism
-
datasetName
Registered dataset identifier (fromFROM <dataset>), ornullfor inlineEXTERNAL. Consumed by the operator factory's per-file_indexsynthesizer so the column carries the user-facing dataset name rather than the resource path. -
deferredExtraction
Whether the plan pairs this source with anExternalFieldExtractExecconsuming deferred-encoded columns. The operator factory keys deferred extraction off this flag, not off_rowPositionpresence in the projection — the latter is also produced for plain_idcomposition with no extract operator downstream. -
declaredReadSpec
The declared mapping's read-instructions (renames,_id.path), orDeclaredReadSpec.NONE. Consumed byFileSourceFactory: renames physicalize reader-facing names,_id.pathstamps_idfrom that column. -
build
-