di-agent-query-optimization

SQL pushdown optimization moves data transformation logic directly into the source database via a single SQL statement instead of splitting processing across systems. This Skill rewrites Substrait plans into a compact full-pushdown form so DataStage can execute more work inside the database.

Substrait plan pushdown classification checks whether every relation and read can be preserved in a single SQL statement with same-connection eligibility. If the plan cannot be fully preserved, the Skill stops rewriting and prevents partial pushdown execution.

You need to provide valid Substrait JSON, explicit connection metadata, and canonical output aliases. The Skill uses these inputs to validate same-connection eligibility and generate an optimized plan for downstream DataStage execution.

Yes, the optimized Substrait plan includes explicit connection metadata, canonical output names, and a trailing SQL SELECT statement. This structure ensures the rewritten full-pushdown plan is ready for immediate downstream DataStage workflow execution.

SQL pushdown optimization stops when a relation cannot be fully preserved to prevent silent data loss or logic splitting across systems. This constraint ensures the rewritten Substrait plan only returns a compact structure when full same-connection SQL execution is verified.