Bkend Data

bkend.ai Data Management

Create tables, define schemas, and perform CRUD operations.

Actions

Creating Tables

Tables in bkend.ai are created through the dashboard or MCP:

Via Dashboard

1. Go to project > Database
2. Click "New Table"
3. Define columns with types

Via MCP

Create a table called "posts" with:
- title (string, required)
- body (text, required)
- status (enum: draft/published, default: draft)
- author_id (reference to users)
- created_at (datetime, auto)
- updated_at (datetime, auto)

Column Types

CRUD Operations

Create

const post = await bkend.data.create('posts', {
  title: 'My First Post',
  body: 'Hello world!',
  status: 'draft',
  author_id: currentUser.id,
});

Read (Single)

const post = await bkend.data.get('posts', postId);

Read (List with Filters)

const posts = await bkend.data.list('posts', {
  status: 'published',
  sort: '-created_at',
  page: '1',
  limit: '20',
});

Update

const updated = await bkend.data.update('posts', postId, {
  title: 'Updated Title',
  status: 'published',
});

Delete

await bkend.data.delete('posts', postId);

Query Patterns

Filtering

// Exact match
{ status: 'published' }

// Multiple filters (AND)
{ status: 'published', author_id: userId }

Sorting

// Ascending
{ sort: 'created_at' }

// Descending (prefix with -)
{ sort: '-created_at' }

Pagination

// Page-based
{ page: '2', limit: '10' }

Search

// Full-text search
{ search: 'keyword' }

React Hooks

// hooks/usePosts.ts
import { useQuery, useMutation, useQueryClient } from '@tanstack/react-query';

export function usePosts(filters?: Record<string, string>) {
  return useQuery({
    queryKey: ['posts', filters],
    queryFn: () => bkend.data.list('posts', filters),
  });
}

export function useCreatePost() {
  const qc = useQueryClient();
  return useMutation({
    mutationFn: (data: CreatePostInput) => bkend.data.create('posts', data),
    onSuccess: () => qc.invalidateQueries({ queryKey: ['posts'] }),
  });
}

Schema Best Practices

1. Always include created_at and updated_at
2. Use reference type for relationships
3. Add indexes on frequently filtered columns
4. Use enum for fixed value sets
5. Use json sparingly (harder to query)

Reference

See references/bkend-patterns.md for complete API client code and advanced patterns.