Introduction

MasterRecord is the ORM half of Master. You describe your tables as plain classes, it generates and runs migrations, and you query with a fluent, lambda-based language — the same code across SQLite, MySQL, and PostgreSQL.

Two styles, one ORM#

MasterRecord supports both the Active Record style (entities save themselves — great for getting started) and the Entity Framework style (a change-tracked context with saveChanges() — efficient for batch work). Use whichever fits the moment.

Code-first models#

Each field is a method that receives the schema builder. MasterRecord reads these to build your tables — no separate schema file, no decorators.

// app/models/Post.js — fields are methods on the class
export default class Post {
  id(db)    { db.integer().primary().auto(); }
  title(db) { db.string().notNullable(); }
  body(db)  { db.text(); }
  views(db) { db.bigint(); }
  created_at(db) { db.datetime(); }
}

An expressive query language#

Queries use real JavaScript lambdas with $$ placeholders for safe, parameterized values — no string concatenation, no injection.

import db from './app/models/db.js';

// Create (Active Record style)
const post = db.Post.new();
post.title = 'By the power of Grayskull';
await post.save();

// Read with a parameterized query
const found = await db.Post
  .where((p) => p.title.like($$), 'By the power%')
  .toList();

// Update
found[0].views = 9000;
await found[0].save();

What’s in the box#

MasterRecord 1.22 deliberately mirrors Entity Framework Core — the EF concept is named next to each feature so you can map your existing knowledge straight across.

• Code-first models with a fluent field builder; TPH inheritance, composite keys, owned JSON types, computed columns, SQL defaults, check constraints, FK constraints with ON DELETE.
• Change tracking as a unit of work — identity map, dirty index, asNoTracking()/asTracking(), entry(), request-scoped contexts and ContextPool (EF DbContext, AddDbContextPool).
• Query language — where, and, like, any, ordering, pagination, aggregates, groupBy().aggregate(), executeUpdate/executeDelete, named global query filters.
• Relationships — belongsTo, hasMany, hasOne, hasManyThrough, manyToMany skip navigations; include / thenInclude / asSplitQuery, explicit and lazy loading.
• Transactions & concurrency — transaction(), savepoints, rowVersion()/concurrencyToken() with ConcurrencyError, retry on transient failures, save events.
• Migrations — schema diffing, atomic apply with a tracking table, script, migrations-status, remove-migration; full CLI reference.
• Transformers, lifecycle hooks, pluggable logging with parameter redaction, health checks.
• Query caching (in-memory or Redis) and full-text search.