Recognises a misfiled data quality dimension
A dimension that does not match what the check measures, so its result rolls up into the wrong column of the quality report and the column it belongs to reads better than it is.
Can AI do this?
People still lead Few or none of the tested AI models showed they master this. 0 of 4 models in the current benchmark master it.
| Model | Tier | Probability of mastery | Status |
|---|---|---|---|
| gemini-4-argon | Flagship | 34% | Not shown |
| gemini-3.1-pro-preview* | Flagship | 20% | Not shown |
| gpt-6-astra* | Flagship | 74% | Not shown |
| claude-fable-5-1* | Flagship | 19% | Not shown |
| gemini-3.8-flash* | Mid tier | 20% | Not shown |
| gpt-6.1-sol* | Mid tier | 73% | Not shown |
| claude-sonnet-5-5* | Mid tier | 71% | Not shown |
| gemini-3.5-flash-lite | Low cost | 37% | Not shown |
| gpt-6-luna | Low cost | 82% | Not shown |
| claude-haiku-5-5 | Low cost | 82% | Not shown |
* From the earlier benchmark of 2026-10: 10 data-modelling test tasks, 3 runs per task, graded by claude-opus-5-5. These models have not taken the current 123 tasks yet; the two methods differ, so compare their figures with care. They are left out of the “most models” counts and the best-model picks.
Mastered means a probability of at least 90% that the model has this knobit, estimated from its graded answers. “Not shown” means the answers do not settle it either way. How the models are tested.
What it looks like in practice
Expected result: The check and the dimension it measures
How it is assessed: Flags the deficient case and passes the acceptable one. Deficient: A uniqueness check filed as completeness; a reference-table lookup filed as accuracy (accuracy needs a source of truth, not a code list) Acceptable: Dimension from the rule's subtype — the table in 07-dq-rules.md (DQR-009, DQR-022)
Also known as: Misfiled dimension (data quality rule)
Related knobits
Builds on
All 30 knobits of Derive data quality rules.
Source: Review checklist 03-dq-rules (data-quality), row 18; DAMA UK 2013; DMBOK2; formerly data-modelling K-718