Skilirium

Recognises a length taken from sample data

K-817 · Skill knobit · Activity: Generate and review DDL · Competence: Data modelling

A VARCHAR/CHAR sized from the data that happened to be loaded in development. An IBAN is up to 34 characters, a BIC is 8 or 11, a SEPA party name is 70. Too short truncates on some engines and errors on others, and a truncated IBAN is a payment to nobody.

Can AI do this?

AI does this partly Some of the tested AI models master this; others do not show it. 3 of 10 models master it.

ModelTierProbability of masteryStatus
gemini-4-argonFlagship 99%Mastered
gemini-3.1-pro-previewFlagship 30%Not shown
gpt-6-astraFlagship 41%Not shown
claude-fable-5-1Flagship 99%Mastered
gemini-3.8-flashMid tier 28%Not shown
gpt-6.1-solMid tier 38%Not shown
claude-sonnet-5-5Mid tier 99%Mastered
gemini-3.5-flash-liteLow cost 28%Not shown
gpt-6-lunaLow cost 39%Not shown
claude-haiku-5-5Low cost 30%Not shown

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 column and the length its domain standard sets

How it is assessed: Flags the deficient case and passes the acceptable one. Deficient: dim_account.account_number VARCHAR(20) on a mart that will also hold non-Estonian accounts: correct for every Estonian IBAN and 14 characters short of a Maltese one Acceptable: Either VARCHAR(34) from the standard with the 20-character Estonian form as a DQ rule scoped to Estonian IBANs, or VARCHAR(20) with the Estonia-only scope written in the column comment and in the rule. What is not acceptable is a length nobody chose

Also known as: Length taken from the sample, not the domain (generated DDL)

Related knobits

Builds on

All 33 knobits of Generate and review DDL.

Source: Review checklist 08-ddl, row 17; ISO 13616; this skill