In Laravel, migrations are not easy to write: you always have to keep in mind what is the current structure of your database to apply changes. And after the migration, you have to deal with the creation and alignment of the models with your changes: you basically have to repeat, again and again, stuff that you have already defined in your migrations (relations, casts, ...).
For me, this was unacceptable, considering how fast you can develop using Laravel: I always though that migrations were the bottleneck of Laravel.
Now, imagine an auto-migration method, that allows you to easily define the structure of your database, and when you change something, it automatically makes the migrations and align the database to your clear-to-read structure. Not only that, but it also generates your models starting from that...
Prisma is a Node.JS ORM and includes Prisma Migrate, which uses Prisma schema (*.schema files) changes to automatically generate database schema migrations. You can read more about it in its official site.
Obviously, Laravel is not a Node.JS environment, and Laravel already provides us a very good ORM, Eloquent. But we still can use Prisma Migrate to easily manage our database.
And via this generator, we can also generate automatically your Eloquent models, filled with PHPDocs, attributes, casts, validation rules, ..., directly from your Prisma schema.
From | To |
---|---|
/// trait:Illuminate\Auth\Authenticatable
/// trait:Illuminate\Database\Eloquent\Factories\HasFactory
/// trait:Illuminate\Notifications\Notifiable
/// implements:Illuminate\Contracts\Auth\Authenticatable as AuthenticatableContract
model User {
id Int @id @default(autoincrement()) /// guarded
email String @unique /// read-only
password String /// hidden, guarded
first_name String?
last_name String?
role Role @default(GUEST)
created_at DateTime @default(now())
updated_at DateTime @default(now())
deleted_at DateTime?
posts Post[]
@@map("users")
} |
<?php
namespace App\Models\Prisma;
use App\Enums\Prisma\Role;
use App\Models\Post;
use Carbon\CarbonImmutable;
use Closure;
use Illuminate\Auth\Authenticatable;
use Illuminate\Contracts\Auth\Authenticatable as AuthenticatableContract;
use Illuminate\Database\Eloquent\Builder;
use Illuminate\Database\Eloquent\Collection;
use Illuminate\Database\Eloquent\Factories\HasFactory;
use Illuminate\Database\Eloquent\Model;
use Illuminate\Database\Eloquent\SoftDeletes;
use Illuminate\Notifications\Notifiable;
use Illuminate\Validation\Rule;
use Illuminate\Validation\Rules\Enum;
/**
* PrismaUser Model.
*
* @mixin Builder
*
* @method static Builder|static query()
* @method static static make(array $attributes = [])
* @method static static create(array $attributes = [])
* @method static static forceCreate(array $attributes)
* @method static firstOrNew(array $attributes = [], array $values = [])
* @method static firstOrFail($columns = ['*'])
* @method static firstOrCreate(array $attributes, array $values = [])
* @method static firstOr($columns = ['*'], Closure $callback = null)
* @method static firstWhere($column, $operator = null, $value = null, $boolean = 'and')
* @method static updateOrCreate(array $attributes, array $values = [])
* @method null|static first($columns = ['*'])
* @method static static findOrFail($id, $columns = ['*'])
* @method static static findOrNew($id, $columns = ['*'])
* @method static null|static find($id, $columns = ['*'])
*
* @property-read int $id
* @property-read string $email
* @property string $password
* @property ?string $first_name
* @property ?string $last_name
* @property Role $role
* @property-read CarbonImmutable $created_at
* @property-read CarbonImmutable $updated_at
* @property-read ?CarbonImmutable $deleted_at
* @property-read Collection<Post> $posts
*/
abstract class PrismaUser extends Model implements AuthenticatableContract
{
use Authenticatable;
use HasFactory;
use Notifiable;
use SoftDeletes;
protected $table = 'users';
protected $attributes;
protected $guarded = ['id', 'password'];
protected $hidden = ['password'];
protected array $rules;
protected $casts;
public function __construct(array $attributes = [])
{
$this->attributes = [
'role' => Role::GUEST,
...$this->attributes ?? [],
];
$this->rules = [
'id' => ['required', 'numeric', 'integer'],
'email' => [
Rule::unique('users', 'email')->ignore(
$this->getKey(),
$this->getKeyName()
),
'required',
'string',
],
'password' => ['required', 'string'],
'first_name' => ['nullable', 'string'],
'last_name' => ['nullable', 'string'],
'role' => ['required', new Enum(Role::class)],
'created_at' => ['required', 'date'],
'updated_at' => ['required', 'date'],
'deleted_at' => ['nullable', 'date'],
...$this->rules ?? [],
];
$this->casts = [
'id' => 'integer',
'email' => 'string',
'password' => 'string',
'first_name' => 'string',
'last_name' => 'string',
'role' => Role::class,
'created_at' => 'immutable_datetime',
'updated_at' => 'immutable_datetime',
'deleted_at' => 'immutable_datetime',
...$this->casts ?? [],
];
parent::__construct($attributes);
}
public function posts()
{
return $this->hasMany(Post::class, 'user_id', 'id');
}
} |
enum Role {
GUEST @map("1")
MODERATOR @map("2")
ADMIN @map("3")
} |
<?php
namespace App\Enums\Prisma;
enum Role: string
{
case GUEST = '1';
case MODERATOR = '2';
case ADMIN = '3';
} |
See a more detailed example in ./packages/usage of this repository. |
In this section we're going to follow you in the setup of prisma-laravel-generator
inside your Laravel project, which can be a brand-new project but also an already existing (and published) project.
Please, follow the instructions carefully and make a backup of your code and your database before proceeding. We take no responsibility for any possible loss of data.
Install prisma
and prisma-laravel-generator
as Node dev-dependencies:
# With Laravel Sail
sail npm install --save-dev prisma @datomatic/prisma-laravel-generator
# NPM
npm install --save-dev prisma @datomatic/prisma-laravel-generator
# Yarn
yarn add -D prisma @datomatic/prisma-laravel-generator
Then, add laravel-prisma-bridge
package to your Composer dependencies:
# With Laravel Sail
sail composer install datomatic/laravel-prisma-bridge
# Without Laravel Sail
composer install datomatic/laravel-prisma-bridge
To work, Prisma needs a Shadow Database, which is used internally as a temporary database.
If you provide to Prisma a user with root privileges, Prisma will create the database by its own.
Otherwise, you have to create a new database (eventually inside the same MySQL instance in which you're running your main database, but it can also be in an entirely different host). In terms of security, this second alternative is, obviously, more secure and preferable. Before proceeding, create a new database that will be used as the shadow database in the next steps. This is not covered by this readme since it's dependent on your environment, except if you're using Laravel Sail with the default mysql image.
IMPORTANT: this section assumes that you're using mysql/mysql-server:8.0
as an image, but you might use a different image and/or database. In that case, you have to manage the creation of the database by yourself.
MORE IMPORTANT: by following this section, you're going to delete your current database. All your data will be lost. Make a dump of your database before proceeding, and restore it when finished.
If you're using Laravel Sail and its default MySQL image, in order to create a new database add to your docker-compose.yml
this volume to your mysql image:
# ...
<your_mysql_container_name>:
image: 'mysql/mysql-server:8.0'
# ...
volumes:
# ...
- './vendor/datomatic/laravel-prisma-bridge/docker-scripts/create-prisma-database.sh:/docker-entrypoint-initdb.d/11-create-prisma-database.sh'
# ...
# ...
Then stop, if it's running, your Sail instance (by executing sail down
or CTRL+C if your in attached mode), and run:
docker rm -v <your_mysql_container_name>
Replace <your_mysql_container_name> with your mysql container name (usually mysql
).
In this way, we're deleting the previous volume (allocated disk space) in order to re-initialize it on the next start, executing the new script.
If the command fails because there isn't a volume with that specific name, execute docker volume ls
and try to find the exact name of the volume of your container, and re-execute the command using the correct name.
Start docker again with sail up
.
If you prefer, instead of doing this, you can add another container (using the same image of the main database container) that initializes another database, but this requires you to understand how Docker Compose works and how to set up the container, and it's not covered by this readme.
When done, set the following variables to your Laravel .env
file (follow the comments):
# You should already have these (otherwise add and set them properly):
DB_CONNECTION=mysql
DB_HOST=localhost
DB_PORT=3306
DB_DATABASE=your_database
DB_USERNAME=username
DB_PASSWORD=*******
# Add and set these accordingly to your settings for the shadow database
DB_SHADOW_CONNECTION=mysql
DB_SHADOW_HOST=localhost
DB_SHADOW_PORT=3306
DB_SHADOW_DATABASE=your_shadow_database
DB_SHADOW_USERNAME=username
DB_SHADOW_PASSWORD=*******
# Add these:
DB_URL="${DB_CONNECTION}://${DB_USERNAME}:${DB_PASSWORD}@${DB_HOST}:${DB_PORT}/${DB_DATABASE}"
DB_SHADOW_URL="${DB_SHADOW_CONNECTION}://${DB_SHADOW_USERNAME}:${DB_SHADOW_PASSWORD}@${DB_SHADOW_HOST}:${DB_SHADOW_PORT}/${DB_SHADOW_DATABASE}"
If you're using the same host and user for both your main database and your shadow database, you can skip some of the DB_SHADOW_***
variables and use DB_***
instead inside of DB_SHADOW_URL
:
# You should already have these (otherwise add and set them properly):
DB_CONNECTION=mysql
DB_HOST=localhost
DB_PORT=3306
DB_DATABASE=your_database
DB_USERNAME=username
DB_PASSWORD=*******
# Add and set this accordingly to your settings for the shadow database
DB_SHADOW_DATABASE=your_shadow_database
# Add these:
DB_URL="${DB_CONNECTION}://${DB_USERNAME}:${DB_PASSWORD}@${DB_HOST}:${DB_PORT}/${DB_DATABASE}"
DB_SHADOW_URL="${DB_CONNECTION}://${DB_USERNAME}:${DB_PASSWORD}@${DB_HOST}:${DB_PORT}/${DB_SHADOW_DATABASE}"
In fact, only the two variables DB_URL
and DB_SHADOW_URL
are going to be used, the other DB_SHADOW_***
variables are used just to keep the configuration cleaner.
Create a new prisma
folder in the root of your Laravel project, create a schema.prisma
file inside, and set the generator
and datasource
sections as following:
generator laravel_generator {
provider = "node node_modules/prisma-laravel-generator"
output = "../"
}
datasource db {
provider = "mysql"
url = env("DB_URL")
shadowDatabaseUrl = env("DB_SHADOW_URL")
}
You can change the generator settings as you prefer: see the Generator Settings section to see all the available settings.
Set the provider of your datasource accordingly to your setup. See Data Sources for more information about the datasource section.
It's important that you execute ALL of your Laravel migrations (also the ones that are executed for the initialization of the DB) after installing laravel-prisma-bridge
.
So, you have to drop all the tables from your database: your database must be completely empty in order to execute all the migrations again.
But, before, you might like the idea to make a dump of your data.
If your development database already has some data in it that you want to keep, make a dump of your data before dropping the tables. Otherwise, jump to the next section Drop all tables.
To make a dump of your data (without structure) in MySQL, you can execute the following command:
mysqldump -u [user] -p [database-name] > dump.sql
If you're using a different type of database, you need to find how to make a dump of the data in your environment.
You're now ready to drop all the tables from your database.
In MySQL the easiest way to drop all of your tables is to delete the database and recreate it:
DROP DATABASE <database>;
CREATE DATABASE <database>;
Replace <database>
with the name of your database.
If you're using a different database, you should find something similar appropriate for your environment.
In order to execute all the initial Laravel migrations, execute the following command:
# With Laravel Sail:
sail artisan migrate
# Without Laravel Sail:
php artisan migrate
Thanks to laravel-prisma-bridge
, this command will now:
- Create the migrations files compatible with Prisma inside your
prisma/migrations
folder; - Execute them via Prisma;
- Log inside the Laravel
migrations
table the fact that the migration was applied; - Sync your
schema.prisma
with the new changes; - Generate updated Laravel models;
If you previously created a dump of your database, you can now restore it.
In MySQL, you can use the command:
mysql -u [user] -p [database] < dump.sql
Take attention that this operation must not delete the _prisma_migrations
table that Prisma created in the previous step.
If you've manually performed some changes in your db in the past (outside of Laravel migrations), after restoring the dump you have to baseline your database. Otherwise, jump to the next section Baselining your production environment.
To baseline your database, execute the following commands in order to create a new migration and then ignore it (since your database is already aligned to that migration):
# With Laravel Sail:
sail npx prisma db pull
sail npx prisma migrate dev --name initial_migration --create-only
sail npx prisma migrate resolve --applied ***_initial_migration
# Without Laravel Sail:
npx prisma db pull
npx prisma migrate dev --name initial_migration --create-only
npx prisma migrate resolve --applied ***_initial_migration
Replace ***_initial_migration
with the name of the migration that Prisma creates (replace ***
with the timestamp).
For more information about what is happening, see Baselining a database.
If you already have a production environment, in order to avoid any data loss you have to set all the migrations that we've just created as already applied: we don't need to execute them, since we did not change (yet) the structure of the database from before.
Obviously, if your production database has some manual changes (outside of Laravel and Prisma migrations), you should first make some migrations in your development environment to align the structure of your development database as your production database before proceeding.
Then, assuming that your database has exactly the same structure of your development database, you have to bring the prisma
folder with all its migrations to it (make a commit, or do whatever you're usually doing to bring updates to your production environment).
You have then to execute the following command in the root of your Laravel project in order to resolve as applied all Prisma migrations:
# With Laravel Sail:
find prisma/migrations -maxdepth 1 -mindepth 1 -type d -print0 | xargs -0 basename -a | xargs -I {} bash -c 'vendor/bin/sail npx prisma migrate resolve --applied "{}"'
# Without Laravel Sail:
find prisma/migrations -maxdepth 1 -mindepth 1 -type d -print0 | xargs -0 basename -a | xargs -I {} bash -c 'npx prisma migrate resolve --applied "{}"'
In this way, we're telling Prisma that all the migrations found in prisma/migrations
are already applied.
If the command (tested on macOS 12.2.1 with zsh 5.8) does not work in your environment, you should write your own command that takes every migration in your prisma/migrations
folder and execute the following command:
# With Laravel Sail:
sail npx prisma migrate resolve --applied [migration-name]
# Without Laravel Sail:
npx prisma migrate resolve --applied [migration-name]
(or you can do it manually instead of writing a command)
For more information about what is happening, see Baselining a database.
You can now play with your schema.prisma
file and use Prisma as you would in a javascript project. See the Usage section for more details.
In this section we're going to see some common stuff regarding the integration of Prisma inside of Laravel.
This readme don't cover how to use Prisma, we suggest you to read the official documentation.
See Prisma documentation:
Thanks to prisma-laravel-generator
, Prisma will generate corresponding models to your tables. See Schema Configuration to see how to influence the generated models directly from your schema.prisma
.
The generated models are filled with PHPDoc, relation methods, casts, validation rules, etc., and are ready-to-use.
By default, the models are generated inside your ./app/Models/Prisma
folder, and prefixed by Prisma
in their name.
The models are stored as abstract classes, and in your ./app/Models
folder you'll find other models without Prisma
prefix, that inherits from them.
You should never touch the models in your ./app/Models/Prisma
folder! Those files are recreated at every generation and any manual change will be lost. If you need to add some methods, properties, etc., change the model in your ./app/Models
, which will not be edited ever again by prisma-laravel-generator
.
When you delete a model from your schema, the related Prisma model in ./app/Models/Prisma
will be deleted, but not the main model in ./app/Models
, which needs to be removed manually. This behaviour is intentional, since the model can contain custom code.
Differently, enums are not generated as "abstract", since in PHP there is no abstraction in enums and there is no inheritance either. Even if they don't have Prisma*
prefix in their names, you should not make changes to the enums, since they're going to be re-generated entirely.
The folders in which the models are created, and how to manipulate the generated model, is further explained in the Schema Configuration section.
- After a
(php|sail) artisan migrate
command (Laravel migration), Prisma will update automatically your models; - After a
(sail) npx prisma migrate dev
command, Prisma will update automatically your models (except if you're using--skip-generate
); - After a
(sail) npx prisma db push
command, Prisma will update automatically your models (except if you're using--skip-generate
);
So, in general, you should never need to update manually your models. But, if you need it, you can execute:
# With Laravel Sail:
sail npx prisma generate
# Without Laravel Sail:
npx prisma generate
In order to execute Laravel migrations, execute the following command:
# With Laravel Sail:
sail artisan migrate
# Without Laravel Sail:
php artisan migrate
Thanks to laravel-prisma-bridge
, this command will now:
- Create the migrations files compatible with Prisma inside your
prisma/migrations
folder; - Execute them via Prisma;
- Log inside the Laravel
migrations
table the fact that the migration was applied; - Sync your
schema.prisma
with the new changes; - Generate updated Laravel models;
You can also roll back a Laravel migration with:
# With Laravel Sail:
sail artisan migrate:rollback
# Without Laravel Sail:
php artisan migrate:rollback
In fact, all the artisan migrate:*
commands are available and integrated with Prisma.
BUT, there is a limit to it: Laravel's migrations should be only used to alter the structure of the tables, NOT to alter the data. When generating the .sql files compatible with Prisma starting from a Laravel migration, just the structural changes are converted.
If you need to alter the data (e.g., you were using Eloquent to do some changes in a migration), you have to find some alternative ways (create an artisan command to migrate the data, or customize Prisma migration to add some manipulations using raw SQL, ...).
It is sometimes useful to squash either some or all migration files into a single migration. See Squashing migrations for more information.
In the generated Prisma*
models you'll find a property called $rules
, which is populated with standard Laravel validation methods for every attribute of your model. This property is not used by Laravel, you have to manage your own validation depending on your APIs. You can create subsets of the rules depending on which attributes you're going to validate, and so on.
See the official Laravel documentation to learn how to use those validation rules.
In this section we're going to see how to affect the generated models and the available configurations for the generator.
Req. | Name | Description | Default |
---|---|---|---|
YES | provider | The generator path. Generally, you don't have to change that. | "node node_modules/@datomatic/prisma-laravel-generator" |
YES | output | The path to the root of your Laravel project. Can be relative (starting from the prisma folder in your project) or absolute. Generally, you don't have to change that. |
"../" |
prismaModelsPath | Path of the generated Prisma* abstract models, starting from the root of your Laravel project. WARNING: This folder will be emptied when the generator runs, so you should set it as a dedicated folder for generated Models. |
"./app/Models/Prisma" |
|
prismaModelsPrefix | The prefix used for the generated abstract models | "Prisma" |
|
prismaEnumsPath | Path of the generated Prisma* enums, starting from the root of your Laravel project. WARNING: This folder will be emptied when the generator runs, so you should set it as a dedicated folder for generated Enums. |
"./app/Enums/Prisma" |
|
modelsPath | Path of the one-time generated models (the one that inherits from the generated Prisma* models which you should use and you are free to change), starting from the root of your Laravel project. |
"./app/Models" |
|
baseModel | FQCN of the class that all the models inherit from | "Illuminate\\Database\\Eloquent\\Model" |
|
basePivotModel | FQCN of the class that all the generated pivots inherit from | "Illuminate\\Database\\Eloquent\\Relations\\Pivot" |
|
phpCsFixerBinPath | Path of PHPCSFixer, starting from the root of your Laravel project (usually, "./vendor/bin/php-cs-fixer" ). If provided, the generated models will be formatted by PHPCSFixer to align the code style of the generated models to the rest of your project. If not assigned, the models will be formatted using Prettier PHP Plugin. |
undefined |
|
phpCsFixerConfigPath | Path of PHPCSFixer config file, starting from the root of your Laravel project (usually, "./.php-cs.dist.php" ). Used if a valid phpCsFixerBinPath is provided. |
undefined |
The name of models and enums can be:
- plural snake_case: The table name will keep that format, while the model name will be converted to singular PascalCase (e.g.
user_categories
toUserCategory
) - singular PascalCase: In that case, the model and the table keep the same name
The conversion from singular to plural (and vice-versa) used by this generator is exactly the same as the one used in Laravel.
There are several annotations that you can add in your schema file to fine tune the generated models and enums. Some annotations are related to an entire model or enum, while others are specific to a field.
In general, all the generator-specific annotations are defined starting by three slashes, i.e. <target> \\\ annotation
.
You can add set multiple annotations to the same target by separating them via comma, i.e., <target> \\\ annotation, another, more
, or by separating them in different rows:
\\\ annotation
\\\ another
\\\ more
<target>
Annotations are case-insensitive. Some annotations are alias of some other attributes available in Prisma.
In models, you can use the following annotations. In this list, we're also covering some of the most important Prisma attributes (the ones that start with @
). To better understand how to manage models in Prisma and to check all the available attributes in Prisma, look at the official documentation.
Define the name of the table in the database. The model will still use the name of the model.
model Example {
id Int @id
@@map("example_table")
}
If you need to set some unique rules that considers multiple columns, you can use this attribute.
model Example {
id Int @id @default(autoincrement())
firstname String
lastname String
@@unique([firstname, lastname])
}
If you need to skip the generation of a model for a specific table, use this attribute
model Example {
id Int @id @default(autoincrement())
@@ignore()
}
Set the model fields as mass-assignable.
/// mass-assignable
model Example {
id Int @id @default(autoincrement())
}
Adds a specific trait to the model. You can also set an alias to the class.
/// trait:App/Traits/ModelTrait
/// trait:App/Traits/AnotherModelTrait as Alias
model Example {
id Int @id @default(autoincrement())
}
Set which class should the model extends. Only one extends:*
can be applied per model (no multiple inheritance is allowed in PHP). You can also set an alias to the class.
/// extends:App/Models/CustomModel
model Example {
id Int @id @default(autoincrement())
}
/// extends:App/Models/CustomModel as Alias
model ExampleWithAlias {
id Int @id @default(autoincrement())
}
Set which interfaces should the model implement. You can also set an alias to the class.
/// implements:App/Contracts/CustomInterface
/// implements:App/Contracts/AnotherCustomInterface as Alias
model Example {
id Int @id @default(autoincrement())
}
In explicit many-to-many relations, you should annotate the pivot model with pivot
.
/// pivot
model Example {
id Int @id @default(autoincrement())
A_id Int
A A @relation(fields: [A_id], references: [id])
B_id Int
B B @relation(fields: [B_id], references: [id])
}
If you need to export a specific model to a different folder to the one defined in your generator settings as modelsPath
(see generator settings),you can use this annotation. The path is relative to the output
path defined in your generator settings.
/// path:./app/Models/domain
model Example {
id Int @id @default(autoincrement())
}
There are several annotations available for fields. In this list, we're also covering some of the most important Prisma attributes (the ones that start with @
). To better understand how to manage fields in Prisma and to check all the available attributes in Prisma, look at the official documentation.
Use the field as primary key.
model Example {
id Int @id
}
Set the default value of the field. This will be set both as database default value but also in laravel $attributes
property.
model Example {
id Int @id @default(autoincrement())
age Int @default(18)
uuid String @default(uuid())
now String @default(now())
}
Use the field as unique. Appropriate validation rules will be created.
model Example {
email String @unique
}
If you need to ignore some attributes in the generated model, use this annotation.
model Example {
id Int @id @default(autoincrement())
interal String @ignore
}
Currently, mapping a field name to a column with a different name is not allowed in Laravel, except for the primary key.
If you try to add @map
to a field that is not the primary key of the model, the generator will throw an error.
model Example {
id Int @id @default(autoincrement()) @map("primary_id")
}
By default, if your field's name is created_at
or updated_at
, you don't need to add these annotations since they're automatically classified as (created/updated)-at timestamp and considered read-only.
If you define a created_at
, you have to define also an updated_at
field (and vice-versa).
You might use @updatedAt
instead of /// updated_at
, if you prefer.
Fields with these annotations are automatically considered as read-only.
model Example {
id Int @id @default(autoincrement())
created_timestamp DateTime /// created_at
updated_timestamp DateTime /// updated_at
}
By default, if your field's name is exactly deleted_at
, you don't need to add these annotations since they're automatically classified as deleted-at timestamp and considered read-only.
When deleted_at
is defined, the model will automatically implement the SoftDelete
trait.
Fields with this annotation are automatically considered as read-only.
model Example {
id Int @id @default(autoincrement())
deleted_timestamp DateTime /// deleted_at
}
Add the field to the model's $guarded
attribute. Can't be used together with fillable
.
model Example {
id Int @id @default(autoincrement())
email String /// guarded
}
Add the field to the model's $fillable
attribute. Can't be used together with guarded
.
model Example {
id Int @id @default(autoincrement())
name String /// fillable
}
Hidden: hidden
Add the field to the model's $hidden
attribute. Can't be used together with visible
.
model Example {
id Int @id @default(autoincrement())
password String /// hidden
}
Add the field to the model's $visible
attribute. Can't be used together with hidden
.
model Example {
id Int @id @default(autoincrement())
email String /// visible
}
Set the field as read-only (via PHPDocs). If the field is a DateTime, instead of a Carbon
instance, it will be casted to CarbonImmutable
.
By default, fields called created_at
, updated_at
and deleted_at
are considered as read-only.
The primary key of your model is automatically considered read-only if the model is not mass-assignable and the field is not fillable/is guarded.
model Example {
id Int @id @default(autoincrement())
email String /// read-only
}
Add the relation to the model's $with
attribute, in order to perform eager-loading.
model Example {
id Int @id @default(autoincrement())
related OtherModel? /// eager-loaded
}
Add the relation to the model's $touches
attribute, in order to update the timestamps of the related model on model changes.
model Example {
id Int @id @default(autoincrement())
related OtherModel? /// touch
}
Both Backed and Pure enums are compatible. To better understand how to manage enums in Prisma, look at the official documentation.
enum Role {
USER
ADMIN
}
enum PostType {
article @map("1")
release @map("2")
news @map("3")
event @map("4")
}
Adds a specific trait to the enum. You can also set an alias to the class.
/// trait:App/Traits/RoleEnumTrait
/// trait:App/Traits/AnotherEnumTrait as Alias
enum Enum {
A
B
}
If you need to export a specific enum to a different folder to the one defined in your generator settings as prismaEnumsPath
(see generator settings),you can use this annotation. The path is relative to the output
path defined in your generator settings.
Warning: If you remove an enum with a custom path from your schema, the generated PHP file has to be deleted manually.
/// path:./app/Enums/domain
enum Enum {
A
B
}
Here you can find some examples on how to define different type of relations. If you need more information about what is happening, look at the official documentation.
Polymorphic relations are not implemented yet. In the meanwhile, you can define the columns in your schema, and define the polymorphic relationship manually in the model.
model A {
id Int @id @default(autoincrement())
related B?
}
model B {
id Int @id @default(autoincrement())
relatedId Int @unique
related A @relation(fields: [relatedId], references: [id])
}
You need to define a relation name to avoid ambiguities. See the official documentation for more details.
model Example {
id Int @id @default(autoincrement())
relatedId Int? @unique
parent Example? @relation("relation_name", fields: [relatedId], references: [id])
child Example? @relation("relation_name")
}
model A {
id Int @id @default(autoincrement())
related B[]
}
model B {
id Int @id @default(autoincrement())
relatedId Int
one_to_one_a2 A @relation(fields: [relatedId], references: [id])
}
You need to define a relation name to avoid ambiguities. See the official documentation for more details.
model Example {
id Int @id @default(autoincrement())
relatedId Int?
parent Example? @relation("relation_name", fields: [relatedId], references: [id])
childs Example[] @relation("relation_name")
}
model A {
id Int @id @default(autoincrement())
related B[]
}
model B {
id Int @id @default(autoincrement())
related A[]
}
model A {
id Int @id @default(autoincrement())
related Pivot[]
}
/// pivot
model Pivot {
id Int @id @default(autoincrement())
A_id Int
A A @relation(fields: [A_id], references: [id])
B_id Int
B B @relation(fields: [B_id], references: [id])
}
model B {
id Int @id @default(autoincrement())
related Pivot[]
}
Some functionalities might not be supported: some are not supported by Prisma, others are not supported by Laravel, some are mis-understanding, and the remaining ones are going to be implemented in future versions!
-
Polymorphic relations: Not implemented yet! In the meanwhile, you can define the columns in your schema, and define the polymorphic relationship manually in the model
-
Databases different from MySQL: Currently MySQL is the only fully-tested driver for this generator, but it should work also with other databases types! We'll try soon to focus also on other databases, to check what works and what not.
-
Pivot with additional relations: If your pivot needs additional relations with other models, you should remove the
pivot
annotation and use it as a standard model. Pivots generation works only if there are just two one-to-many relations. -
Allow data alterations inside Laravel migrations: This is not possible. Laravel's migrations should be only used to alter the structure of the tables, NOT to alter the data. When generating the .sql files compatible with Prisma starting from a Laravel migration, just the structural changes are converted. If you need to alter the data (e.g., you were using Eloquent to do some changes in a migration), you have to find some alternative ways (create an artisan command to migrate the data, or customize Prisma migration to add some manipulations using raw SQL, ...).
-
Composite primary keys: Not supported by Laravel
-
created_at without updated_at (and vice-versa): Not supported by Laravel
-
cuid(): Not supported natively by Laravel (you can use some external packages, but this is up to you)
-
Mapped fields (e.g.,
firstName String @map("first_name")
): Laravel does not provide a way to map attributes to columns with different names, except for the primary key. -
Mapped enums (backed enums) only allows string as db-value (i.e., you can't set
@map(1)
but only@map("1")
): This is limited by Prisma, since @map allows as an argument only strings -
fillable
andguarded
fields at the same time: Does not make sense in Laravel -
hidden
andvisible
fields at the same time: Does not make sense in Laravel -
mass-assignable
withguarded
orfillable
fields:mass-assignable
annotation adds aprotected $guarded = [];
attribute to the model, so it does not make sense to define other fields as guarded or fillable. -
Multiple data sources: Not supported by Prisma