strapi/docs/3.0.0-beta.x/concepts/models.md

# Models

## Concept

Models are a representation of the database's structure and lifecycle. They are split into two separate files. A JavaScript file that contains the lifecycle callbacks, and a JSON one that represents the data stored in the database and their format. The models also allow you to define the relationships between them.

**Path —** `./api/restaurant/models/Restaurant.js`.

```js
module.exports = {
  // Before saving a value.
  // Fired before an `insert` or `update` query.
  beforeSave: (model, attrs, options) => {},

  // After saving a value.
  // Fired after an `insert` or `update` query.
  afterSave: (model, attrs, options) => {},

  // ... and more
};
```

**Path —** `./api/restaurant/models/Restaurant.settings.json`.

```json
{
  "connection": "default",
  "info": {
    "name": "restaurant",
    "description": "This represents the Restaurant Model"
  },
  "attributes": {
    "cover": {
      "collection": "file",
      "via": "related",
      "plugin": "upload"
    },
    "name": {
      "default": "",
      "type": "string"
    },
    "description": {
      "default": "",
      "type": "text"
    }
  }
}
```

In this example, there is a `Restaurant` model which contains two attributes `cover`, `name` and `description`.

### Where are the models defined?

The models are defined in each `./api/**/models/` folder. Every JavaScript or JSON file in these folders will be loaded as a model. They are also available through the `strapi.models` and `strapi.api.**.models` global variables. Usable everywhere in the project, they contain the ORM model object that they are refer to. By convention, models' names should be written in lowercase.

## How to create a model?

::: tip
If you are just starting out it is very convenient to generate some models with the Content Type Builder, directly in the admin interface. You can then review the generated model mappings on the code level. The UI takes over a lot of validation tasks and gives you a fast feeling for available features.
:::

Use the CLI, and run the following command `strapi generate:model restaurant name:string description:text`. Read the [CLI documentation](../cli/CLI.md) for more information.

This will create two files located at `./api/restaurant/models`:

- `Restaurant.settings.json`: contains the list of attributes and settings. The JSON format makes the file easily editable.
- `Restaurant.js`: imports `Restaurant.settings.json` and extends it with additional settings and lifecycle callbacks.

::: tip
when you create a new API using the CLI (`strapi generate:api <name>`), a model is automatically created.
:::

## Model settings

Additional settings can be set on models:

- `connection` (string) - Connection's name which must be used. Default value: `default`.
- `collectionName` (string) - Collection's name (or table's name) in which the data should be stored.
- `globalId` (string) -Global variable name for this model (case-sensitive).

**Path —** `Restaurant.settings.json`.

```json
{
  "connection": "mongo",
  "collectionName": "Restaurants_v1",
  "globalId": "Restaurants",
  "attributes": {}
}
```

In this example, the model `Restaurant` will be accessible through the `Restaurants` global variable. The data will be stored in the `Restaurants_v1` collection or table and the model will use the `mongo` connection defined in `./config/environments/**/database.json`

::: tip
The `connection` value can be changed whenever you want, but you should be aware that there is no automatic data migration process. Also if the new connection doesn't use the same ORM you will have to rewrite your queries.
:::

## Model information

The info key on the model-json states information about the model. This information is used in the admin interface, when showing the model.

- `name`: The name of the model, as shown in admin interface.
- `description`: The description of the model.
- `mainField`: Determines which model-attribute is shown when displaying the model.

**Path —** `Restaurant.settings.json`.

```json
{
  "info": {
    "name": "restaurant",
    "description": ""
  }
}
```

## Model options

The options key on the model-json states.

- `idAttribute`: This tells the model which attribute to expect as the unique identifier for each database row (typically an auto-incrementing primary key named 'id'). _Only valid for strapi-hook-bookshelf_
- `idAttributeType`: Data type of `idAttribute`, accepted list of value below. _Only valid for strapi-hook-bookshelf_
- `timestamps`: This tells the model which attributes to use for timestamps. Accepts either `boolean` or `Array` of strings where first element is create date and second element is update date. Default value when set to `true` for Bookshelf is `["created_at", "updated_at"]` and for MongoDB is `["createdAt", "updatedAt"]`.
- `uuid` : Boolean to enable UUID support on MySQL, you will need to set the `idAttributeType` to `uuid` as well and install the `bookshelf-uuid` package. To load the package you can see [this example](./configurations.md#bookshelf-mongoose).

**Path —** `User.settings.json`.

```json
{
  "options": {
    "timestamps": true
  }
}
```

## Define the attributes

The following types are currently available:

- `string`
- `text`
- `richtext`
- `integer`
- `biginteger`
- `float`
- `decimal`
- `password`
- `date`
- `time`
- `datetime`
- `timestamp`
- `boolean`
- `binary`
- `uuid`
- `enumeration`
- `json`
- `email`

### Validations

You can apply basic validations to the attributes. The following supported validations are _only supported by MongoDB_ connection.
If you're using SQL databases, you should use the native SQL constraints to apply them.

- `required` (boolean) — if true adds a required validator for this property.
- `unique` (boolean) — whether to define a unique index on this property.
- `index` (boolean) — adds an index on this property, this will create a [single field index](https://docs.mongodb.com/manual/indexes/#single-field) that will run in the background (_only supported by MongoDB_).
- `max` (integer) — checks if the value is greater than or equal to the given minimum.
- `min` (integer) — checks if the value is less than or equal to the given maximum.

**Security validations**
To improve the Developer eXperience when developing or using the administration panel, the framework enhances the attributes with these "security validations":

- `private` (boolean) — if true, the attribute will be removed from the server response (it's useful to hide sensitive data).
- `configurable` (boolean) - if false, the attribute isn't configurable from the Content Type Builder plugin.

### Example

**Path —** `Restaurant.settings.json`.

```json
{
  ...
  "attributes": {
    "title": {
      "type": "string",
      "min": 3,
      "max": 99,
      "unique": true
    },
    "description": {
      "default": "My descrioption",
      "type": "text",
      "required": true
    },
    ...
  }
}
```

## Relations

Relations let your create links (relations) between your Content Types.

:::: tabs

::: tab "One-Way" id="one-way"

One-way relationships are useful to link an entry to another. However, only one of the models can be queried with its populated items.

#### Example

A `pet` can be owned by someone (a `user`).

**Path —** `./api/pet/models/Pet.settings.json`.

```json
{
  "attributes": {
    "owner": {
      "model": "user"
    }
  }
}
```

**Example**

```js
// Create a pet
const xhr = new XMLHttpRequest();
xhr.open('POST', '/pets', true);
xhr.setRequestHeader('Content-Type', 'application/json');
xhr.send(
  JSON.stringify({
    owner: '5c151d9d5b1d55194d3209be', // The id of the user you want to link
  })
);
```

:::

::: tab "One-to-One" id="one-to-one"

One-to-One relationships are usefull when you have one entity that could be linked to only one other entity. And vis versa.

#### Example

A `user` can have one `address`. And this address is only related to this `user`.

**Path —** `./api/user/models/User.settings.json`.

```json
{
  "attributes": {
    "address": {
      "model": "address",
      "via": "user"
    }
  }
}
```

**Path —** `./api/address/models/Address.settings.json`.

```json
{
  "attributes": {
    "user": {
      "model": "user"
    }
  }
}
```

**Example**

```js
// Create an address
const xhr = new XMLHttpRequest();
xhr.open('POST', '/addresses', true);
xhr.setRequestHeader('Content-Type', 'application/json');
xhr.send(
  JSON.stringify({
    user: '5c151d9d5b1d55194d3209be', // The id of the user you want to link
  })
);
```

:::

::: tab "One-to-Many" id="one-to-many"

One-to-Many relationships are usefull when an entry can be liked to multiple entries of an other Content Type. And an entry of the other Content Type can be linked to only one entry.

#### Example

A `user` can have many `articles`, and an `article` can be related to one `user` (author).

**Path —** `./api/user/models/User.settings.json`.

```json
{
  "attributes": {
    "articles": {
      "collection": "article",
      "via": "author"
    }
  }
}
```

**Path —** `./api/article/models/Article.settings.json`.

```json
{
  "attributes": {
    "author": {
      "model": "user"
    }
  }
}
```

**Examples**

```js
// Create an article
const xhr = new XMLHttpRequest();
xhr.open('POST', '/articles', true);
xhr.setRequestHeader('Content-Type', 'application/json');
xhr.send(
  JSON.stringify({
    author: '5c151d9d5b1d55194d3209be', // The id of the user you want to link
  })
);

// Update an article
const xhr = new XMLHttpRequest();
xhr.open('PUT', '/users/5c151d9d5b1d55194d3209be', true);
xhr.setRequestHeader('Content-Type', 'application/json');
xhr.send(
  JSON.stringify({
    articles: ['5c151d51eb28fd19457189f6', '5c151d51eb28fd19457189f8'], // Set of ALL articles linked to the user (existing articles + new article or - removed article)
  })
);
```

:::

::: tab "Many-to-Many" id="many-to-many"

One-to-Many relationships are usefull when an entry can be liked to multiple entries of an other Content Type. And an entry of the other Content Type can be linked to many entries.

#### Example

A `product` can be related to many `categories`, so a `category` can have many `products`.

**Path —** `./api/product/models/Product.settings.json`.

```json
{
  "attributes": {
    "categories": {
      "collection": "category",
      "via": "products",
      "dominant": true
    }
  }
}
```

**NOTE**:
(NoSQL databases only) The `dominant` key defines which table/collection should store the array that defines the relationship. Because there are no join tables in NoSQL, this key is required for NoSQL databases (ex: MongoDB).

**Path —** `./api/category/models/Category.settings.json`.

```json
{
  "attributes": {
    "products": {
      "collection": "product",
      "via": "categories"
    }
  }
}
```

**Example**

```js
// Update a product
const xhr = new XMLHttpRequest();
xhr.open('PUT', '/products/5c151d9d5b1d55194d3209be', true);
xhr.setRequestHeader('Content-Type', 'application/json');
xhr.send(
  JSON.stringify({
    categories: ['5c151d51eb28fd19457189f6', '5c151d51eb28fd19457189f8'], // Set of ALL categories linked to the product (existing categories + new category or - removed category)
  })
);
```

:::

::: tab "Polymorphic" id="polymorphic"

The polymorphic relationships are the solution when you don't know which kind of model will be associated to your entry. A common use case is an `Image` model that can be associated to many others kind of models (Article, Product, User, etc).

#### Single vs Many

Let's stay with our `Image` model which might belongs to **a single `Article` or `Product` entry**.

**NOTE**:
In other words, it means that an `Image` entry can be associated to one entry. This entry can be a `Article` or `Product` entry.

**Path —** `./api/image/models/Image.settings.json`.

```json
{
  "attributes": {
    "related": {
      "model": "*",
      "filter": "field"
    }
  }
}
```

Also, our `Image` model which might belongs to **many `Article` or `Product` entries**.

**NOTE**:
In other words, it means that an `Article` entry can relate to the same image as a `Product` entry.

**Path —** `./api/image/models/Image.settings.json`.

```json
{
  "attributes": {
    "related": {
      "collection": "*",
      "filter": "field"
    }
  }
}
```

#### Filter

The `filter` attribute is optional (but we highly recommend to use it every time). If it's provided it adds a new match level to retrieve the related data.

For example, the `Product` model might have two attributes which are associated to the `Image` model. To distinguish which image is attached to the `cover` field and which images are attached to the `pictures` field, we need to save and provide this to the database.

**Path —** `./api/article/models/Product.settings.json`.

```json
{
  "attributes": {
    "cover": {
      "model": "image",
      "via": "related"
    },
    "pictures": {
      "collection": "image",
      "via": "related"
    }
  }
}
```

The value is the `filter` attribute is the name of the column where the information is stored.

#### Example

A `Image` model might belongs to many either `Article` models or a `Product` models.

**Path —** `./api/image/models/Image.settings.json`.

```json
{
  "attributes": {
    "related": {
      "collection": "*",
      "filter": "field"
    }
  }
}
```

**Path —** `./api/article/models/Article.settings.json`.

```json
{
  "attributes": {
    "avatar": {
      "model": "image",
      "via": "related"
    }
  }
}
```

**Path —** `./api/article/models/Product.settings.json`.

```json
{
  "attributes": {
    "pictures": {
      "collection": "image",
      "via": "related"
    }
  }
}
```

#### Database implementation

If you're using MongoDB as a database, you don't need to do anything. Everything is natively handled by Strapi. However, to implement a polymorphic relationship with SQL databases, you need to create two tables.

**Path —** `./api/image/models/Image.settings.json`.

```json
{
  "attributes": {
    "name": {
      "type": "string"
    },
    "url": {
      "type": "string"
    },
    "related": {
      "collection": "*",
      "filter": "field"
    }
  }
}
```

The first table to create is the table which has the same name as your model.

```
CREATE TABLE `image` (
  `id` int(11) NOT NULL,
  `name` text NOT NULL,
  `text` text NOT NULL
)
```

**NOTE**:
If you've overridden the default table name given by Strapi by using the `collectionName` attribute. Use the value set in the `collectionName` to name the table.

The second table will allow us to associate one or many others entries to the `Image` model. The name of the table is the same as the previous one with the suffix `_morph`.

```
CREATE TABLE `image_morph` (
  `id` int(11) NOT NULL,
  `image_id` int(11) NOT NULL,
  `related_id` int(11) NOT NULL,
  `related_type` text NOT NULL,
  `field` text NOT NULL
)
```

- `image_id` is using the name of the first table with the suffix `_id`.
  - **Attempted value:** It corresponds to the id of an `Image` entry.
- `related_id` is using the attribute name where the relation happens with the suffix `_id`.
  - **Attempted value:** It corresponds to the id of an `Article` or `Product` entry.
- `related_type` is using the attribute name where the relation happens with the suffix `_type`.
  - **Attempted value:** It corresponds to the table name where the `Article` or `Product` entry is stored.
- `field` is using the filter property value defined in the model. If you change the filter value, you have to change the name of this column as well.
  - **Attempted value:** It corresponds to the attribute of an `Article`, `Product` with which the `Image` entry is related.

| id  | image_id | related_id | related_type | field  |
| --- | -------- | ---------- | ------------ | ------ |
| 1   | 1738     | 39         | product      | cover  |
| 2   | 4738     | 58         | article      | avatar |
| 3   | 1738     | 71         | article      | avatar |

:::

::::

## Lifecycle callbacks

::: warning
The life cycle functions are based on the ORM life cycle and not on the strapi request.
We are currently working on it to make it easier to use and understand.
Please check [this issue](https://github.com/strapi/strapi/issues/1443) on GitHub.
:::

The following events are available by default:

Callbacks on:

:::: tabs

::: tab save

`save`

- beforeSave
- afterSave

:::

::: tab fetch

`fetch`

- beforeFetch
- afterFetch

::: tab fetchAll

`fetchAll`

- beforeFetchAll
- afterFetchAll

:::

::: tab create

`create`

- beforeCreate
- afterCreate

:::

::: tab update

`update`

- beforeUpdate
- afterUpdate

:::

::: tab destroy

`destroy`

- beforeDestroy
- afterDestroy

:::

::::

### Example

:::: tabs

::: tab Mongoose

#### Mongoose

The entry is available through the `model` parameter

**Path —** `./api/user/models/User.js`.

```js
module.exports = {
  /**
   * Triggered before user creation.
   */
  beforeCreate: async model => {
    // Hash password.
    const passwordHashed = await strapi.api.user.services.user.hashPassword(
      model.password
    );

    // Set the password.
    model.password = passwordHashed;
  },
};
```

:::

::: tab Bookshelf

#### Bookshelf

Each of these functions receives a three parameters `model`, `attrs` and `options`. You have to return a Promise.

**Path —** `./api/user/models/User.js`.

```js
module.exports = {
  /**
   * Triggered before user creation.
   */
  beforeCreate: async (model, attrs, options) => {
    // Hash password.
    const passwordHashed = await strapi.api.user.services.user.hashPassword(
      model.attributes.password
    );

    // Set the password.
    model.set('password', passwordHashed);
  },
};
```

:::

:::::