Skip to main content
Version: Next


By default, all entities are monkey patched with toObject() and toJSON methods:

interface AnyEntity<K = number | string> {
toObject(parent?: AnyEntity, isCollection?: boolean): Record<string, any>;
toJSON(...args: any[]): Record<string, any>;
// ...

When you serialize your entity via JSON.stringify(entity), its toJSON method will be called automatically. You can provide custom implementation for toJSON, while using toObject for initial serialization:

class Book {

// ...

toJSON(strict = true, strip = ['id', 'email'], ...args: any[]): { [p: string]: any } {
const o = wrap(this, true).toObject(...args); // do not forget to pass rest params here

if (strict) {
strip.forEach(k => delete o[k]);

return o;


Do not forget to pass rest params when calling toObject(...args), otherwise the results might not be stable.

Hidden Properties

If you want to omit some properties from serialized result, you can mark them with hidden flag on @Property() decorator. To have this information available on the type level, you also need to use the HiddenProps symbol:

class Book {

// we use the `HiddenProps` symbol to define hidden properties on type level
[HiddenProps]?: 'hiddenField' | 'otherHiddenField';

@Property({ hidden: true })
hiddenField =;

@Property({ hidden: true, nullable: true })
otherHiddenField?: string;


const book = new Book(...);
console.log(wrap(book).toObject().hiddenField); // undefined

// @ts-expect-error accessing `hiddenField` will fail to compile thanks to the `HiddenProps` symbol
console.log(wrap(book).toJSON().hiddenField); // undefined

Alternatively, you can use the Hidden type. It works the same as the Opt type (an alternative for OptionalProps symbol), and can be used in two ways:

  • with generics: hiddenField?: Hidden<string>;
  • with intersections: hiddenField?: string & Hidden;

Both will work the same, and can be combined with the HiddenProps symbol approach.

class Book {

@Property({ hidden: true })
hiddenField: Hidden<Date> =;

@Property({ hidden: true, nullable: true })
otherHiddenField?: string & Hidden;


Shadow Properties

The opposite situation where you want to define a property that lives only in memory (is not persisted into database) can be solved by defining your property as persist: false. Such property can be assigned via one of Entity.assign(), em.create() and em.merge(). It will be also part of serialized result.

This can be handled when dealing with additional values selected via QueryBuilder or MongoDB's aggregations.

class Book {

@Property({ persist: false })
count?: number;


const book = new Book(...);
wrap(book).assign({ count: 123 });
console.log(wrap(book).toObject().count); // 123
console.log(wrap(book).toJSON().count); // 123

Property Serializers

As an alternative to custom toJSON() method, we can also use property serializers. They allow to specify a callback that will be used when serializing a property:

class Book {

@ManyToOne({ serializer: value =>, serializedName: 'authorName' })
author: Author;


const author = new Author('God')
const book = new Book(author);
console.log(wrap(book).toJSON().authorName); // 'God'

Implicit serialization

Implicit serialization means calling toObject() or toJSON() on the entity, as opposed to explicitly using the serialize() helper. Since v6, it works entirely based on populate hints. This means that, unless you explicitly marked some entity as populated via wrap(entity).populated(), it will be part of the serialized form only if it was part of the populate hint:

// let's say both Author and Book entity has a m:1 relation to Publisher entity
// we only populate the publisher relation of the Book entity
const user = await em.findOneOrFail(Author, 1, {
populate: ['books.publisher'],

const dto = wrap(user).toObject();
console.log(dto.publisher); // only the FK, e.g. `123`
console.log(dto.books[0].publisher); // populated, e.g. `{ id: 123, name: '...' }`

Moreover, the implicit serialization now respects the partial loading hints too. Previously, all loaded properties were serialized, partial loading worked only on the database query level. Since v6, we also prune the data on runtime. This means that unless the property is part of the partial loading hint (fields option), it won't be part of the DTO. Main difference here is the primary and foreign keys, that are often automatically selected as they are needed to build the entity graph, but will no longer be part of the DTO.

const user = await em.findOneOrFail(Author, 1, {
fields: [''],

const dto = wrap(user).toObject();
// only the publisher's name will be available + primary keys
// `{ id: 1, books: [{ id: 2, publisher: { id: 3, name: '...' } }] }`

This also works for embeddables, including nesting and object mode.

Primary keys are automatically included. If you want to hide them, you have two options:

  • use hidden: true in the property options
  • use serialization: { includePrimaryKeys: false } in the ORM config

Foreign keys are forceObject

Unpopulated relations are serialized as foreign key values, e.g. { author: 1 }, if you want to enforce objects, e.g. { author: { id: 1 } }, use serialization: { forceObject: false } in your ORM config.

For strict typings to respect the global config option, you need to define it on your entity class via Config symbol:

import { Config, Entity, ManyToOne, PrimaryKey, Ref, wrap } from '@mikro-orm/core';

class Book {

[Config]?: DefineConfig<{ forceObject: true }>;

id!: number;

@ManyToOne(() => User, { ref: true })
author!: Ref<User>;


const book = await em.findOneOrFail(Book, 1);
const dto = wrap(book).toObject();
const identityId =; // without the Config symbol, `dto.identity` would resolve to number

Explicit serialization

The serialization process is normally driven by the populate hints. If you want to take control over this, you can use the serialize() helper:

import { serialize } from '@mikro-orm/core';

const dtos = serialize([user1, user2]);
// [
// { name: '...', books: [1, 2, 3], identity: 123 },
// { name: '...', ... },
// ]

const [dto] = serialize(user1); // always returns an array
// { name: '...', books: [1, 2, 3], identity: 123 }

// for a single entity instance we can as well use `wrap(e).serialize()`
const dto2 = wrap(user1).serialize();
// { name: '...', books: [1, 2, 3], identity: 123 }

By default, every relation is considered as not populated - this will result in the foreign key values to be present. Loaded collections will be represented as arrays of the foreign keys. To control the shape of the serialized response we can use the second options parameter:

interface SerializeOptions<T extends object, P extends string = never, E extends string = never> {
/** Specify which relation should be serialized as populated and which as a FK. */
populate?: AutoPath<T, P>[] | boolean;

/** Specify which properties should be omitted. */
exclude?: AutoPath<T, E>[];

/** Enforce unpopulated references to be returned as objects, e.g. `{ author: { id: 1 } }` instead of `{ author: 1 }`. */
forceObject?: boolean;

/** Ignore custom property serializers. */
ignoreSerializers?: boolean;

/** Skip properties with `null` value. */
skipNull?: boolean;

/** Only include properties for a specific group. If a property does not specify any group, it will be included, otherwise only properties with a matching group are included. */
groups?: string[];

Here is a more complex example:

import { wrap } from '@mikro-orm/core';

const dto = wrap(author).serialize({
populate: ['', 'books.publisher', 'favouriteBook'], // populate some relations
exclude: [''], // skip property of some relation
forceObject: true, // not populated or not initialized relations will result in object, e.g. `{ author: { id: 1 } }`
skipNull: true, // properties with `null` value won't be part of the result

If you try to populate a relation that is not initialized, it will have same effect as the forceObject option - the value will be represented as object with just the primary key available.

Serialization groups

Every property can specify its serialization groups, which are then used with explicit serialization.

Properties without the groups option are always included.

Let's consider the following entity:

class User {

id!: number;

username!: string;

@Property({ groups: ['public', 'private'] })
name!: string;

@Property({ groups: ['private'] })
email!: string;


Now when you call serialize():

  • without the groups option, you get all the properties
  • with groups: ['public'] you get id, username and name properties
  • with groups: ['private'] you get id, username, name and email properties
  • with groups: [] you get only the id and username properties (those without groups)
const dto1 = serialize(user);
// User { id: 1, username: 'foo', name: 'Jon', email: '' }

const dto2 = serialize(user, { groups: ['public'] });
// User { id: 1, username: 'foo', name: 'Jon' }