Entity Framework Core Best Practices

Use when configuring entities, optimizing queries, or implementing multi-tenant data access with EF Core.

When to use

• Configuring entity mappings, indexes, relationships, and value objects
• Optimizing read queries (AsNoTracking, projections) or preventing N+1 problems
• Adding row version concurrency tokens or handling conflicts
• Applying global tenant filters or implementing repositories

Core principles

• Prefer Fluent API over data annotations; keep entities clean with private setters
• Use AsNoTracking for reads; Include/split queries to avoid N+1; project to DTOs when possible
• Index tenant columns and frequent filter combinations; create unique composites for business rules
• Add RowVersion for optimistic concurrency; catch DbUpdateConcurrencyException
• Apply global query filters for multi-tenancy; bypass with IgnoreQueryFilters when needed
• Encapsulate in repositories or inject DbContext directly into services

Resources

• Entity configuration: [patterns/entity-configuration.md](patterns/entity-configuration.md)
• Value objects: [patterns/value-objects.md](patterns/value-objects.md)
• Query performance: [patterns/query-performance.md](patterns/query-performance.md)
• Indexing strategy: [patterns/indexing.md](patterns/indexing.md)
• Optimistic concurrency: [patterns/concurrency.md](patterns/concurrency.md)
• Global query filters: [patterns/query-filters.md](patterns/query-filters.md)
• Repository pattern: [patterns/repository-pattern.md](patterns/repository-pattern.md)

Default locations

• Entity configurations: src/GloboTicket.Infrastructure/Persistence/Configurations
• DbContext: src/GloboTicket.Infrastructure/Persistence/GloboTicketDbContext.cs
• Repositories (if used): src/GloboTicket.Infrastructure/Persistence/Repositories
• Migrations: src/GloboTicket.Infrastructure/Migrations

Validation checklist

• Entities have private setters and parameterless constructors for EF
• Configurations use Fluent API with separate IEntityTypeConfiguration classes
• Tenant columns indexed; composite indexes match query patterns
• Read-only queries use AsNoTracking and project to DTOs
• Include/split queries prevent N+1; large collections use AsSplitQuery
• Concurrency tokens configured where updates are common
• Global filters applied to ITenantEntity types; IgnoreQueryFilters used sparingly
• SaveChanges called in repositories or services consistently