Skip to content

Build a V3-native query engine #247

@em3s

Description

@em3s

Background

v2.engine.Graph is a god class – early tech debt.
The V3 API path currently runs on V2Backed* adapters over V2 internals – a transitional shape.
This issue replaces the query path with a V3-native engine. Mutation is out of scope (do it later).

Task

  • Engine skeleton – Empty Engine Spring bean with a catalog reload loop
  • Catalog reload – Engine periodically loads databases / tables / aliases from the metastore via a CatalogLoader
  • Query – Engine produces TableBinding instances; QueryService uses it

Done When

  • V3 query controllers serve traffic through Engine instead of V2BackedTableBinding.
  • Graph is unchanged.
  • Each phase lands as one or more independently shippable PRs.

Notes

  • "Catalog" = the registry of Database / Table / Alias definitions (Spark / Trino / Iceberg / Glue use the same term). The metastore stays as the Actionbase Table backing store; the catalog is what lives in memory and is loaded from the metastore.
  • Reuse core.edge.mapper.*, core.edge.payload.*, core.metadata.common.* — these are already V3.

Metadata

Metadata

Assignees

Labels

enhancementNew feature or request

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions