[SPARK-28426] Metadata Handling in Thrift Server

• Type: Umbrella
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.0.0
• Fix Version/s: 3.0.0
• Component/s: SQL
• Target Version/s: 3.0.0
• Assignee: Yuming Wang
• Reporter: Xiao Li
• Created: 17/Jul/19 16:45
• Updated: 09/Oct/20 20:33
• Resolved: 20/Aug/19 05:05

Description

Currently, only the `executeStatement is handled` for SQL commands. But others like `getTables`, `getSchemas`, `getColumns` and so on fallback to an in-memory derby empty catalog. Then some of BI tools could not show the correct object information. 

 

This umbrella Jira is tracking the related improvement of Thrift-server. 

Issue Links

• is duplicated by: SPARK-9686 Spark Thrift server doesn't return correct JDBC metadata

Sub-Tasks

• Implement Spark's own GetColumnsOperation: Resolved
• Spark Thrift Server - SQL Client connections does't show db artefacts: Resolved
• SparkSQL - show schemas/tables in dropdowns of SQL client tools (ie Squirrel SQL, DBVisualizer.etc): Resolved
• Show metadata operations on ThriftServerTab: Resolved
• Avoid creating new sessions in SparkMetadataOperationSuite: Resolved
• Show global temporary view in database tool: Resolved
• Implement Spark's own GetTableTypesOperation: Resolved
• Implement Spark's own GetFunctionsOperation: Resolved
• Add CLOSED state to ExecutionState: Resolved
• ThriftServerTab lost error message: Resolved
• Implement GetCatalogsOperation for Thrift Server: Resolved
• Support ThriftServer GetTypeInfoOperation for Spark's own type: Resolved