Database schema diagrams

Best for “what does the data look like?” — a relational schema drawn as an entity-relationship diagram. It’s Mermaid’s erDiagram: a customer places orders; each order contains order_items, and every line item points at a product. ORDER_ITEM is the join table that resolves the many-to-many between orders and products.

erDiagram
  CUSTOMER ||--o{ ORDER : places
  ORDER ||--|{ ORDER_ITEM : contains
  PRODUCT ||--o{ ORDER_ITEM : ordered_as

  CUSTOMER {
    bigint id PK
    text email UK "lowercased on write"
    text name
    timestamptz created_at
  }
  ORDER {
    bigint id PK
    bigint customer_id FK
    text status "pending | paid | shipped | cancelled"
    numeric total
    timestamptz created_at
  }
  PRODUCT {
    bigint id PK
    text sku UK
    text name
    numeric price
    int stock
  }
  ORDER_ITEM {
    bigint id PK
    bigint order_id FK
    bigint product_id FK
    int quantity
    numeric unit_price
  }

The first keyword is erDiagram.

• Relationships read A <card>--<card> B : label — the crow’s-foot cardinality markers (|| one, o{ zero-or-many, |{ one-or-many) sit at each end, and the label names the relationship.
• Each ENTITY { … } block lists columns as type name [key]. Column tags render as badges — PK, FK, UK — and a quoted string after a column becomes an info note.

This is the same web-shop domain modelled as an object graph and a lifecycle on the Class & State page. To tint a table, use a vizzy-fence style override — Mermaid’s own style/classDef only apply inside flowcharts.