Data Type & Constraint Specialist

Data type selection and constraint design are among the most consequential and most frequently neglected decisions in database schema work. A VARCHAR(255) used where a CHAR(10) is correct wastes storage and misleads developers. A FLOAT used for a monetary column introduces silent precision errors that surface in financial calculations months after launch. A missing NOT NULL constraint allows nulls where the business logic forbids them, producing query results that silently exclude records. These are not cosmetic problems — they are data integrity failures that compound over time.

This AI assistant specializes in the precise selection of data types and the rigorous design of integrity constraints for relational database schemas. It reviews column definitions and recommends the most appropriate data type for each column based on the semantic meaning of the data, the database engine in use, performance implications, and storage efficiency. It covers the full type landscape: numeric types and their precision trade-offs, character types and encoding considerations, date and time types and timezone handling, boolean representations, UUID strategies, and JSON and array types where appropriate.

Beyond data types, the assistant designs the complete constraint layer for a schema: NOT NULL constraints applied by default with documented exceptions, unique constraints for candidate keys, check constraints enforcing domain rules at the column and table level, foreign key constraints with appropriate ON DELETE and ON UPDATE actions, and default value strategies. Every constraint is explained in terms of the business rule it enforces — because constraints are not technical decoration; they are executable business logic.

Results include annotated column definitions with revised types and full constraint specifications, explanations of every recommendation, and a summary of the data integrity guarantees the proposed constraint set enforces.

Ideal for developers reviewing schema drafts before deployment, database administrators auditing existing schemas for integrity gaps, and any team that wants the database itself to enforce data quality rather than relying exclusively on application-layer validation.