Executable Data Quality (Real SQL, 6 Dimensions)
Quality stops being an opinion: rules that execute real SQL against the source, in 8 dialects.
Framework / standard: ISO 8000 · DAMA-DMBOK · ISO/IEC 38505-1 (Quality)
Data quality that is not measured is an opinion. Linedat turns quality into rules that execute real SQL against the source — not a simulation that "should" pass in production.
The rules cover the six DAMA/ISO 8000 dimensions (completeness, uniqueness, validity, consistency, timeliness and accuracy), with a catalogue of templates so you do not start from a blank page.
Real SQL in 8 dialects, with templates
The engine generates SQL portable to eight certified engines (PostgreSQL, MySQL, Snowflake, BigQuery, Oracle, SQL Server, DB2 and Databricks) and starts from more than 20 templates across the six dimensions, with concrete validations such as IBAN structure (ISO 13616) or NIF. If no connector is available, execution is recorded as an error — no result is fabricated.
Weighted score and profiling from the source
Each rule contributes to a per-dimension score, weighted by severity. And to design rules with real data rather than blind guesses, column profiling brings nulls, cardinality, min/max, mean/median/standard deviation and top values directly from the source, with sampling. You detect "age" with negative values before writing a single rule.
The limits (what we do not claim)
There is no automatic ML-based anomaly detection: it is rule plus threshold. Executing a rule does not by itself recalculate the aggregate score of the asset (it must be recalculated separately). Regular expression support varies by dialect (for example, it is limited in SQL Server).
How Linedat helps
Linedat moves your team from "we run rules" to "we demonstrate quality coverage by domain and dimension", with real SQL and per-domain reports.
Related capabilities
Without a contract, every schema change is Russian roulette with someone else's data.
Data Valuation: Utility IndexHow much is your data worth? A self-calculated utility index 0–100 (usage, lineage, quality).
Temporal and Verifiable Lineage (Point-in-Time)We know what the lineage looked like on the fiscal year-end — and we can prove nobody touched it.