ecto-schema-patterns

Ecto associations connect schemas through `has_many`, `belongs_to`, and `many_to_many` macros. Use `many_to_many` with join tables to link resources like products to tags, or define a dedicated join schema to store additional metadata such as user roles within organizations.

Embedded schemas store nested data within a parent schema using `embeds_one` or `embeds_many`, enabling you to model hierarchical or composite data without separate database tables. Use them for unstructured nested fields that belong exclusively to their parent record.

Multi-tenancy in Ecto uses schema prefixes to isolate tenant data across separate database schemas or tables. Define a custom prefix on your schema to route queries to tenant-specific namespaces, ensuring data segregation in applications serving multiple organizations.

Yes, Ecto supports custom primary keys beyond the default auto-incrementing integer. Define a primary key with `:uuid` type or use `:string` for custom identifiers like SKUs, then set `@primary_key {:field_name, :type, []}` in your schema definition.

Virtual fields exist in memory only and are never persisted to the database, useful for computed values or temporary form data. Define them with `field :name, :type, virtual: true` to exclude them from `INSERT` and `UPDATE` operations.

Use `:map` or `{:array, :type}` field types to store unstructured or list data, then apply validation in custom changesets using `cast`, `validate_required`, and custom validation functions to ensure data integrity before persistence.