```json { "table_name": "search_index_metrics", "table_type_tags": [ {"tag": "event", "confidence": 0.80}, {"tag": "path", "confidence": 0.30} ], "primary_type": "event", "insufficient_evidence": false, "entity": { "columns": ["dataset_id", "batch_id", "engine"], "unique_per_row": false }, "time_axis": { "columns": ["created_at"], "kind": "event_time" }, "reasoning": "The composite entity (dataset_id, batch_id, engine) repeats across rows (dataset_id RPUV 4.67, batch_id RPUV 3.02, engine RPUV 148.00), ordered by created_at event timestamps (Uniqueness 0.667). Each row records an index build occurrence — a dataset+batch was indexed by an engine at a specific time, capturing performance metrics (build_ms, index_bytes, docs_indexed). The rows are bare occurrences to count and analyze; replaying them does not reconstruct a meaningful state of the entity. The unique id (Uniqueness 1.000) is an artificial occurrence identifier. Event is the primary reading; path has weak secondary confidence because the rows do track batch processing history, though no cumulative state is derived." } ```