bknd-define-relationship（bknd 定义关系）

Define Entity Relationships

Create relationships between entities in Bknd (foreign keys, references, associations).

Prerequisites

• At least two entities exist (see bknd-create-entity)
• For code mode: Access to your schema file

Relationship Types

When to Use UI vs Code

Use UI Mode When

• Quick prototyping
• Visual learners
• Non-developers setting up relationships

Use Code Mode When

• Version control needed
• Reproducible schema
• Custom options (mappedBy, connectionTable)
• Team collaboration

UI Approach

Step 1: Access Data Section

1. Start server: npx bknd run
2. Open http://localhost:1337
3. Navigate to Data section

Step 2: Add Relation Field

1. Click on the child entity (e.g., posts)
2. Click + Add Field
3. Select Relation field type
4. Choose the target entity (e.g., users)
5. Select relationship type:

- Many-to-One: Multiple posts can belong to one user - One-to-One: One post has exactly one user - Many-to-Many: Posts can have many tags, tags can have many posts

Step 3: Configure Options

• Field Name: Name for the foreign key (e.g., author creates author_id)
• Required: Toggle if relationship is mandatory

Step 4: Save and Sync

1. Click Save Field
2. Click Sync Database to apply changes

Code Approach

Relationships are defined in the second argument to em():

const schema = em(
  {
    // Entity definitions (first argument)
  },
  ({ relation, index }, entities) => {
    // Relationship definitions (second argument)
  }
);

Many-to-One

Child belongs to one parent. Most common relationship type.

import { em, entity, text } from "bknd";

const schema = em(
  {
    users: entity("users", { email: text().required() }),
    posts: entity("posts", { title: text().required() }),
  },
  ({ relation }, { users, posts }) => {
    relation(posts).manyToOne(users);
  }
);

Auto-generated: users_id foreign key column on posts table

Custom field name with mappedBy:

({ relation }, { users, posts }) => {
  relation(posts).manyToOne(users, {
    mappedBy: "author",  // Creates author_id instead of users_id
  });
}

One-to-One

Exclusive 1:1 relationship. Each child belongs to exactly one parent.

const schema = em(
  {
    users: entity("users", { email: text().required() }),
    profiles: entity("profiles", { bio: text() }),
  },
  ({ relation }, { users, profiles }) => {
    relation(profiles).oneToOne(users);
  }
);

Note: One-to-one relationships cannot use $set operator (maintains exclusivity).

Many-to-Many

Both entities can have multiple of the other. Junction table created automatically.

const schema = em(
  {
    posts: entity("posts", { title: text().required() }),
    tags: entity("tags", { name: text().required() }),
  },
  ({ relation }, { posts, tags }) => {
    relation(posts).manyToMany(tags);
  }
);

Auto-generated: posts_tags junction table with posts_id and tags_id columns

Custom junction table name:

({ relation }, { posts, tags }) => {
  relation(posts).manyToMany(tags, {
    connectionTable: "post_tags",  // Custom junction table name
  });
}

Extra fields on junction table:

({ relation }, { users, courses }) => {
  relation(users).manyToMany(courses, {
    connectionTable: "enrollments",
  }, {
    // Extra fields on junction table
    enrolled_at: date(),
    completed: boolean(),
    grade: number(),
  });
}

Self-Referencing

Entity references itself. Common for hierarchies (categories, comments, org charts).

const schema = em(
  {
    categories: entity("categories", { name: text().required() }),
  },
  ({ relation }, { categories }) => {
    relation(categories).manyToOne(categories, {
      mappedBy: "parent",      // FK field: parent_id
      inversedBy: "children",  // Reverse navigation
    });
  }
);

Usage:

• category.parent_id → Points to parent category
• Query children: api.data.readMany("categories", {where: {parent_id: 5}})

Alternative: Direct Foreign Key

Instead of relation(), use .references() on a number field:

const schema = em({
  users: entity("users", { email: text().required() }),
  posts: entity("posts", {
    title: text().required(),
    author_id: number().references("users.id"),
  }),
});

Difference: .references() is simpler but doesn't create inverse navigation or support many-to-many.

Relation Options

ManyToOne / OneToOne Options

ManyToMany Options

Querying Relations

Load Related Data (with)

const api = app.getApi();

// Load posts with their author
const posts = await api.data.readMany("posts", {
  with: {
    users: { select: ["email", "name"] },
  },
});
// Result: [{ id: 1, title: "...", users: { email: "...", name: "..." } }]

Filter by Relation

// Posts by specific author
const posts = await api.data.readMany("posts", {
  where: { author_id: 5 },
});

// Using join for complex filters
const posts = await api.data.readMany("posts", {
  join: {
    users: { where: { email: "john@example.com" } },
  },
});

Many-to-Many Operations

// Attach tags to post
await api.data.updateOne("posts", 1, {
  tags: { $attach: [1, 2, 3] },  // Tag IDs
});

// Detach tags
await api.data.updateOne("posts", 1, {
  tags: { $detach: [2] },
});

// Replace all tags
await api.data.updateOne("posts", 1, {
  tags: { $set: [4, 5] },
});

Many-to-One Operations

// Set author on post
await api.data.updateOne("posts", 1, {
  users: { $set: 5 },  // User ID
});

Common Patterns

Blog with Authors and Tags

const schema = em(
  {
    users: entity("users", {
      email: text().required().unique(),
      name: text(),
    }),
    posts: entity("posts", {
      title: text().required(),
      content: text(),
      published: boolean(),
    }),
    tags: entity("tags", {
      name: text().required().unique(),
    }),
  },
  ({ relation }, { users, posts, tags }) => {
    // Post has one author
    relation(posts).manyToOne(users, { mappedBy: "author" });

    // Posts have many tags
    relation(posts).manyToMany(tags);
  }
);

E-commerce Orders

const schema = em(
  {
    customers: entity("customers", { email: text().required() }),
    orders: entity("orders", { total: number() }),
    products: entity("products", { name: text().required(), price: number() }),
  },
  ({ relation }, { customers, orders, products }) => {
    // Order belongs to customer
    relation(orders).manyToOne(customers);

    // Order has many products (with quantity)
    relation(orders).manyToMany(products, {
      connectionTable: "order_items",
    }, {
      quantity: number().required(),
      unit_price: number().required(),
    });
  }
);

Nested Categories

const schema = em(
  {
    categories: entity("categories", {
      name: text().required(),
      slug: text().required().unique(),
    }),
  },
  ({ relation }, { categories }) => {
    relation(categories).manyToOne(categories, {
      mappedBy: "parent",
      inversedBy: "children",
    });
  }
);

// Usage: Get all children of category 5
const children = await api.data.readMany("categories", {
  where: { parent_id: 5 },
});

Common Pitfalls

Entity Not Found

Error: Entity "user" not found

Fix: Entity names are plural by convention. Use users not user.

// Wrong
relation(posts).manyToOne(user);

// Correct
relation(posts).manyToOne(users);

Circular Reference Error

Error: Circular dependency detected

Fix: For self-referencing, use proper options:

// Correct self-reference
relation(categories).manyToOne(categories, {
  mappedBy: "parent",
  inversedBy: "children",
});

Foreign Key Naming Conflict

Error: Field "users_id" already exists

Fix: Use mappedBy to specify a different field name:

// If you already have users_id, use a different name
relation(posts).manyToOne(users, { mappedBy: "author" });  // Creates author_id

Many-to-Many $set on One-to-One

Error: Cannot use $set on one-to-one relation

Fix: One-to-one maintains exclusivity differently. Use $create instead:

// For one-to-one
await api.data.updateOne("users", 1, {
  profiles: { $create: { bio: "Hello" } },
});

Missing Entity in Destructure

Error: Cannot read property 'manyToOne' of undefined

Fix: Ensure entity is destructured from second callback parameter:

// Wrong - missing users in destructure
({ relation }, { posts }) => {
  relation(posts).manyToOne(users);  // users is undefined
}

// Correct
({ relation }, { users, posts }) => {
  relation(posts).manyToOne(users);
}

Relation Changes Not Applying

Problem: Added relation but not seeing FK column.

Fixes:

1. Restart server (schema syncs on startup)
2. Verify relation is in second em() argument
3. Check for syntax errors

Verification

Check Foreign Key Created

npx bknd debug paths
# Look for the FK field in entity output

Test Relation in Code

const api = app.getApi();

// Create parent
const user = await api.data.createOne("users", { email: "test@example.com" });

// Create child with relation
const post = await api.data.createOne("posts", {
  title: "Test Post",
  author_id: user.data.id,
});

// Load with relation
const loaded = await api.data.readOne("posts", post.data.id, {
  with: { users: true },
});
console.log(loaded.data.users);  // { id: 1, email: "test@example.com" }

DOs and DON'Ts

DO:

• Use plural entity names (users, posts)
• Use mappedBy for semantic field names (author instead of users)
• Define relations in the second em() argument
• Use .references() for simple FK without navigation

DON'T:

• Use singular entity names in relations
• Create manual FK fields when using relation() (it creates them automatically)
• Use $set on one-to-one relations
• Forget to destructure entities in the callback

Related Skills

• bknd-create-entity - Create entities before defining relationships
• bknd-add-field - Add fields including .references() for simple FKs
• bknd-crud-read - Query related data with with and join
• bknd-crud-update - Use $attach, $detach, $set for relation updates
• bknd-query-filter - Advanced filtering on relations