```json { "classifications": [ { "column": "dataset_id", "role": "identifier", "confidence": 1.0, "reason": "Primary key with prefix 'd_' pattern, used to identify dataset rows" }, { "column": "batch_id", "role": "identifier", "confidence": 1.0, "reason": "Foreign key with 'batch_' prefix, unique per row, used for joins" }, { "column": "tenant_id", "role": "dimension", "confidence": 0.95, "reason": "Low cardinality (1 value), used for filtering/grouping by tenant" }, { "column": "submitted_at", "role": "dimension", "confidence": 0.95, "reason": "Timestamp for slicing/filtering data by submission time" }, { "column": "insights_ready_at", "role": "dimension", "confidence": 0.95, "reason": "Timestamp for slicing/filtering by completion time" }, { "column": "time_to_insights_ms", "role": "measure", "confidence": 1.0, "reason": "Numeric duration metric to be aggregated (avg/sum/min/max)" }, { "column": "status", "role": "dimension", "confidence": 1.0, "reason": "Categorical field with 6 distinct values for grouping/filtering" }, { "column": "status_reason", "role": "attribute", "confidence": 0.95, "reason": "Free-form descriptive text explaining status, displayed but not grouped" }, { "column": "phases_json", "role": "attribute", "confidence": 0.9, "reason": "JSON content field (all null), descriptive data not for aggregation" }, { "column": "created_at", "role": "dimension", "confidence": 0.95, "reason": "Timestamp for filtering/grouping by creation time" }, { "column": "updated_at", "role": "dimension", "confidence": 0.95, "reason": "Timestamp for filtering/grouping by update time" }, { "column": "insights_started_at", "role": "dimension", "confidence": 0.95, "reason": "Timestamp for slicing/filtering by start time" } ] } ```