```json { "table_name": "ip_geo", "table_type_tags": [ {"tag": "dimension", "confidence": 0.75}, {"tag": "path", "confidence": 0.30} ], "primary_type": "dimension", "insufficient_evidence": false, "entity": { "columns": ["ip"], "unique_per_row": true }, "time_axis": { "columns": ["resolved_at"], "kind": "event_time" }, "reasoning": "The entity ip is unique per row (Uniqueness 1.000, RPUV 1.00). The table records geolocation lookups for IP addresses — each row describes WHERE an IP resolved to at a point in time. The resolved_at timestamp (Uniqueness 1.000) is the lookup time, not an axis ordering state transitions of the IP itself. The geographic attributes (country, city, lat/lon) are descriptive properties of the lookup result. This is a catalog of IP geolocation records → dimension. A weak path reading exists because resolved_at could order successive lookups of the same IP if IPs were re-resolved over time, but the data shows each IP appears exactly once (ip Uniqueness 1.000), so no entity repeats along the axis." } ```