Consistent naming convention across tables and columns

Inconsistent naming conventions are a signature of AI-generated schemas assembled across multiple sessions without coherence review. A schema mixing userId, post_id, CreatedAt, and updated_at forces developers to look up column names for every query, breaks ORM mappings that assume a consistent convention, and produces mapping errors when tools like Prisma generate TypeScript types. ISO 25010 maintainability-readability captures the downstream cost: every new developer must reverse-engineer the implicit rules before they can write correct queries. Naming inconsistencies also cause real bugs when case-sensitive column lookups fail in production queries.

Pick one convention — PostgreSQL standard is snake_case with plural table names — and apply it via RENAME COLUMN or ORM @map directives. For Prisma codebases that want camelCase TypeScript but snake_case SQL, use @map and @@map.

model User {
  id        String   @id
  createdAt DateTime @map("created_at")
  @@map("users")
}

-- Rename inconsistent columns
ALTER TABLE posts RENAME COLUMN "postTitle" TO title;
ALTER TABLE posts RENAME COLUMN "CreatedAt" TO created_at;

For existing applications with ORM-mapped models, prefer @map over database renames to avoid breaking existing queries until you can migrate them all at once.