# TASK 5.3: Implement Redshift Staging Load Execution, Diagnostics, And Rerun Semantics

## Summary

Implement the execution boundary for the first `load_to_staging` stage: run Redshift `COPY` for the resolved table set, apply the chosen rerun-safe staging replacement strategy, and publish safe per-table diagnostics that later stages can consume.

## Dependencies

- [TASK 5.2](./TASK5.2.md) - staging table catalog and Redshift `COPY` command construction

## Files to Create

### 1. Staging-load executor and diagnostics models

Create or finalize the components that execute staging-load SQL and return table-level diagnostics.

Recommended logical service:

- `RedshiftStagingLoadExecutor`

Required diagnostics details:

- contract name
- target table name
- schema version
- input artifact URIs or prefixes
- timing
- safe status summary
- `copySqlTemplateId` or equivalent execution identifier

## Files to Modify

### 2. Warehouse execution path

Modify the tenant warehouse execution path so `load_to_staging` can run the chosen replacement strategy and then execute `COPY` statements for every required staging table.

### 3. Load-to-staging batch details

Modify or finalize the `load_to_staging` `details_json` shape so it preserves upstream normalize-load references and per-table staging execution results.

## Required Behavior

- the chosen v1 staging replacement strategy is applied consistently before or alongside `COPY` as required
- Redshift `COPY` executes for every required table in the resolved upstream inventory
- if any required table load fails, the endpoint fails and the full `load_to_staging` batch fails
- already written upstream `load/` artifacts remain for replay and diagnostics
- diagnostics must not store secrets
- batch metadata must preserve enough detail for `TASK6.md` to resolve `staging_to_curated` inputs confidently

## Testing

### Unit tests

Add unit tests for:

- execution ordering of replacement strategy and `COPY`
- per-table success and failure diagnostics recording
- failure behavior when one required table load fails
- rerun-safe stage behavior and lifecycle integration
- diagnostics sanitization

## Acceptance Criteria

- [x] Redshift staging load executes for every required table in the resolved upstream inventory
- [x] The chosen replacement strategy is applied consistently for reruns
- [x] Table-load failures fail the endpoint and stage correctly
- [x] Diagnostics stay secret-free and downstream-readable
- [x] Unit tests cover execution, rerun semantics, and diagnostics behavior

## Notes

- This subtask creates the first warehouse-loaded `staging` replay input for curated SQL.
- Keep Redshift execution ownership in shared Warehouse code rather than in the Brink adapter.