Diesel best practices

Use when working with Diesel ORM in Rust services, CLIs, or jobs backed by Postgres, SQLite, or MySQL.

Schema ownership

• Treat schema.rs and the actual database schema as coupled artifacts; regenerate or update them whenever migrations change tables or columns.
• Keep Diesel models close to their table domain and avoid giant files that mix every table, insert struct, and query helper together.
• Use explicit field names and Rust types that mirror nullability and database constraints instead of smoothing over mismatches with ad hoc conversions.

Query builder and raw SQL

• Prefer Diesel’s query builder for common CRUD, filtering, joins, pagination, and updates so backend-specific SQL generation stays type checked.
• Reach for raw SQL only when the query builder becomes meaningfully less clear or cannot express the query cleanly.
• Keep raw SQL isolated, typed, and reviewed carefully; treat it as an exception path, not the default style.

Derives and model shapes