```json { "table_name": "jobs", "table_type_tags": [ {"tag": "dimension", "confidence": 0.95} ], "primary_type": "dimension", "insufficient_evidence": false, "entity": { "columns": ["job_id"], "unique_per_row": true }, "time_axis": { "columns": ["started_at", "finished_at", "created_at", "updated_at"], "kind": "event_time" }, "reasoning": "The entity job_id is unique per row (Uniqueness 1.000, RPUV 1.00), making each row describe one distinct job execution. The timestamps (started_at, finished_at, created_at, updated_at) are attributes of each job — they record when that specific job was created, started, and completed, not a sequence of state changes to a repeating entity. The table is a catalog of job runs: each job_id appears once with its complete execution record (type, status, elapsed_ms, error_count). This is dimension despite the timestamps — the job_id uniqueness is decisive." } ```