async-query-core library

This directory contains async-query-core library, which implements the core logic of async-query and provide extension points to allow plugin different implementation of data storage, etc. async-query module provides implementations for OpenSearch index based implementation.

Type of queries

There are following types of queries, and the type is automatically identified by analysing the query.

• BatchQuery: Execute single query in Spark
• InteractiveQuery: Establish session and execute queries in single Spark session
• IndexDMLQuery: Handles DROP/ALTER/VACUUM operation for Flint indices
• RefreshQuery: One time query request to refresh(update) Flint index
• StreamingQuery: Continuously update flint index in single Spark session

Extension points

Following is the list of extension points where the consumer of the library needs to provide their own implementation.

• Data store interface
  • AsyncQueryJobMetadataStorageService
  • SessionStorageService
  • StatementStorageService
  • FlintIndexMetadataService
  • FlintIndexStateModelService
  • IndexDMLResultStorageService
• Other
  • LeaseManager
  • JobExecutionResponseReader
  • QueryIdProvider
  • SessionIdProvider
  • SessionConfigSupplier
  • EMRServerlessClientFactory
  • SparkExecutionEngineConfigSupplier
  • DataSourceSparkParameterComposer
  • GeneralSparkParameterComposer
  • SparkSubmitParameterModifier To be deprecated in favor of GeneralSparkParameterComposer