nullable but not null

9 months ago

Adding a new field to a model often starts as nullable to avoid table locks during migration.
Application logic is updated to populate the field, followed by a backfill job for existing records.
The final step of making the field non-nullable is often forgotten, leading to schema-data mismatch.
Leaving fields nullable when they shouldn't be creates confusion, missed constraints, and unnecessary complexity.
A script is provided to identify nullable fields that could be made non-nullable by checking actual null percentages.
The script scans models for nullable fields and calculates the percentage of rows with null values.
Fields with 0% nulls should be updated to non-nullable to improve schema accuracy and simplify code.
Example output shows fields with their null percentages, helping identify candidates for schema cleanup.
Fixing these issues ensures the data model is honest, improves validation, and keeps the application healthier.

Hasty Briefsbeta