Unify primary key field as id across MongoDB and SQL drivers

docs/api/README.md

@@ -38,6 +38,48 @@ ObjectQL provides a **unified query protocol** that can be exposed through multi
 4. **Secure**: Built-in validation, permission checks, SQL injection prevention
 5. **Universal**: Same query works across MongoDB, PostgreSQL, SQLite
 
+### Unified ID Field
+
+ObjectQL uses a **unified `id` field** as the primary key across all database drivers:
+
+- **Consistent Naming**: Always use `id` in API requests and responses
+- **Database Agnostic**: Works seamlessly with both MongoDB (which uses `_id` internally) and SQL databases
+- **Automatic Mapping**: MongoDB driver transparently converts between `id` (API) and `_id` (database)
+
+**Example:**
+```json
+// Create with custom ID - works with any driver
+{
+  "op": "create",
+  "object": "users",
+  "args": {
+    "id": "user-123",
+    "name": "Alice"
+  }
+}
+
+// Query by ID - works with any driver
+{
+  "op": "find",
+  "object": "users",
+  "args": {
+    "filters": [["id", "=", "user-123"]]
+  }
+}
+
+// Response always uses 'id'
+{
+  "data": [
+    {
+      "id": "user-123",
+      "name": "Alice"
+    }
+  ]
+}
+```
+
+See the [Driver Documentation](../guide/drivers/index.md) for more details.
+
 ---
 
 ## JSON-RPC Style API

docs/guide/drivers/index.md

@@ -11,6 +11,32 @@ We currently support the following official drivers:
 *   **[SQL Driver (Knex)](./sql)**: Supports PostgreSQL, MySQL, SQLite, MSSQL, etc.
 *   **[MongoDB Driver](./mongo)**: Supports MongoDB.
 
+## Unified ID Field
+
+ObjectQL provides a **consistent API** across all database drivers by standardizing on the `id` field name for primary keys:
+
+- **MongoDB Driver**: Automatically maps `id` (API) ↔ `_id` (database)
+- **SQL Driver**: Uses `id` natively in the database schema
+
+This means you can write database-agnostic code:
+
+```typescript
+// Same code works with MongoDB OR SQL drivers!
+const user = await app.create('users', {
+  id: 'user-123',     // Works with both drivers
+  name: 'Alice'
+});
+
+const query = {
+  filters: [['id', '=', 'user-123']]  // Consistent across drivers
+};
+const results = await app.find('users', query);
+```
+
+**No more switching between `_id` and `id` depending on your database!**
+
+See the individual driver documentation for implementation details.
+
 ## Configuring a Driver
 
 Drivers are instantiated and passed to the `ObjectQL` constructor under the `driver` property (or `datasources` map for multi-db setup).

docs/guide/drivers/mongo.md

@@ -25,10 +25,128 @@ const driver = new MongoDriver({
 });
 ```
 
-## IDs and `_id`
+## Unified ID Field
 
-MongoDB uses `_id` as the primary key. ObjectQL will map the standard `_id` field to the ObjectQL `_id` field.
-When querying, ObjectQL handles the conversion between string IDs and `ObjectId` automatically.
+ObjectQL provides a **unified `id` field** across all database drivers for a consistent API experience.
+
+### How It Works
+
+- **API Level**: You always use `id` in your application code (queries, filters, documents)
+- **Database Level**: MongoDB internally uses `_id` as required by MongoDB conventions
+- **Automatic Mapping**: The driver transparently converts between `id` and `_id`
+
+### Examples
+
+**Querying by ID:**
+```typescript
+// ✅ Use 'id' in queries - works consistently across all drivers
+const query = {
+  filters: [['id', '=', '507f1f77bcf86cd799439011']]
+};
+const users = await app.find('users', query);
+```
+
+**Creating Documents:**
+```typescript
+// ✅ Use 'id' when creating - the driver maps it to '_id' internally
+const newUser = await app.create('users', {
+  id: '507f1f77bcf86cd799439011',  // Optional: specify custom ID
+  name: 'Alice',
+  email: 'alice@example.com'
+});
+
+// Result returned with 'id' field (not '_id')
+console.log(newUser.id); // '507f1f77bcf86cd799439011'
+```
+
+**Finding by ID:**
+```typescript
+// ✅ Use 'id' parameter
+const user = await app.findOne('users', '507f1f77bcf86cd799439011');
+console.log(user.id); // Always 'id', never '_id'
+```
+
+**Sorting by ID:**
+```typescript
+const query = {
+  sort: [['id', 'desc']]  // ✅ Use 'id' for sorting
+};
+const users = await app.find('users', query);
+```
+
+**Field Projection:**
+```typescript
+const query = {
+  fields: ['id', 'name', 'email']  // ✅ Use 'id' to select the ID field
+};
+const users = await app.find('users', query);
+// Results contain 'id', not '_id'
+```
+
+### Migration from Legacy Code
+
+If you have existing code using `_id`, you have two options:
+
+1. **Recommended: Migrate to `id`** for consistency across drivers:
+
+**Before (Legacy):**
+```typescript
+// ❌ Old way - inconsistent with SQL drivers
+const query = {
+  filters: [['_id', '=', '507f1f77bcf86cd799439011']],
+  sort: [['_id', 'desc']],
+  fields: ['_id', 'name']
+};
+```
+
+**After (Recommended):**
+```typescript
+// ✅ New way - consistent across all drivers
+const query = {
+  filters: [['id', '=', '507f1f77bcf86cd799439011']],
+  sort: [['id', 'desc']],
+  fields: ['id', 'name']
+};
+```
+
+2. **Backward Compatible Mode:** Continue using `_id` in queries (results still return `id`)
+
+The MongoDB driver **fully supports `_id` in filters, sorting, and field projections** for backward compatibility:
+
+```typescript
+// ✅ This works - backward compatible
+const query = {
+  filters: [['_id', '=', '507f1f77bcf86cd799439011']],
+  sort: [['_id', 'desc']],
+  fields: ['_id', 'name']
+};
+const users = await app.find('users', query);
+
+// Results ALWAYS use 'id' regardless of query field name
+console.log(users[0].id); // '507f1f77bcf86cd799439011'
+console.log(users[0]._id); // undefined
+```
+
+**Key Points:**
+- Queries accept both `id` and `_id` (automatically mapped to MongoDB's `_id`)
+- Results always return `id` field (never `_id`)
+- Using `id` is recommended for database portability
+
+## ID Generation
+
+When creating documents without specifying an `id`, the driver automatically generates a string ID:
+
+```typescript
+const newUser = await app.create('users', {
+  name: 'Bob',
+  // No id specified
+});
+
+// Driver generates a unique string ID (not ObjectId)
+console.log(newUser.id); // e.g., '507f1f77bcf86cd799439011'
+```
+
+The generated IDs are hexadecimal strings (24 characters) that maintain MongoDB's uniqueness guarantees without using ObjectId objects.
 
 ## Limitations
 

docs/guide/drivers/sql.md

@@ -44,6 +44,74 @@ const driver = new KnexDriver({
 });
 ```
 
+## Unified ID Field
+
+ObjectQL uses **`id`** as the primary key field name across all SQL databases for consistency with the MongoDB driver.
+
+### How It Works
+
+- **Database Column**: SQL tables use `id` as the primary key column (VARCHAR/TEXT type)
+- **API Level**: You use `id` in your queries, filters, and documents
+- **Consistency**: Same API as MongoDB driver - no need to remember `_id` vs `id`
+
+### Examples
+
+**Table Schema:**
+```sql
+CREATE TABLE users (
+  id VARCHAR(255) PRIMARY KEY,  -- Primary key is 'id'
+  name VARCHAR(255),
+  email VARCHAR(255),
+  created_at TIMESTAMP,
+  updated_at TIMESTAMP
+);
+```
+
+**Creating Documents:**
+```typescript
+// Create with auto-generated ID
+const user = await app.create('users', {
+  name: 'Alice',
+  email: 'alice@example.com'
+});
+console.log(user.id); // Auto-generated UUID or custom ID
+
+// Create with custom ID
+const user = await app.create('users', {
+  id: 'custom-user-123',
+  name: 'Bob'
+});
+```
+
+**Querying by ID:**
+```typescript
+const query = {
+  filters: [['id', '=', 'custom-user-123']]
+};
+const users = await app.find('users', query);
+```
+
+**Finding by ID:**
+```typescript
+const user = await app.findOne('users', 'custom-user-123');
+```
+
+### Legacy `_id` Support
+
+For backward compatibility, if you provide `_id` in a create operation, the driver will automatically map it to `id`:
+
+```typescript
+// Legacy code - automatically mapped
+const user = await app.create('users', {
+  _id: 'user-123',  // Mapped to 'id' internally
+  name: 'Charlie'
+});
+
+console.log(user.id); // 'user-123'
+```
+
+This ensures smooth migration for projects that previously used MongoDB-style `_id` fields.
+
 ## Schema Mapping
 
 The driver automatically maps ObjectQL types to SQL column types: