```json { "table_name": "cleandata_v5", "table_type_tags": [ {"tag": "dimension", "confidence": 0.75}, {"tag": "misc", "confidence": 0.35} ], "primary_type": "dimension", "insufficient_evidence": false, "entity": {"columns": ["_unnamed_0"], "unique_per_row": true}, "time_axis": {"columns": [], "kind": null}, "reasoning": "_unnamed_0 has Uniqueness 1.000 (RPUV 1.00), establishing a unique entity per row. However, 11 rows contain placeholder values (NA in f_boxer/s_boxer, the repeated 3.76724137931034 across numeric columns) that corrupt the dataset's semantic coherence. The valid 116 rows describe individual boxing match outcomes with f_boxer vs s_boxer, their attributes (age, height, reach, record), and match result — each row is unique information about one bout. No time axis exists (no event timestamp, no sequence column). The placeholder rows and the lack of clear semantic structure in the entity (an artificial row number rather than a natural match identifier) weaken confidence, but the dominant pattern is dimension: unique rows cataloging match records." } ```