Database Seeding in Laravel

I just discovered an area that might help all developers and someone may have answers already.

With regards to database seeding , it seems the framework is geared around the ORM more so than traditional crud operations with a database.

With that said in seeding a database the framework is a little fuzzy when constraints exist within the database. And the database doesn't seed according to the constraint logic the developer intends.

It is difficult to debug .. is there a way echo the dd() into a dummy browser, outside of a terminal window from the Seeder PHP? My DDs also echo to the terminal.

LaryAI

1 year ago

Level 58

When dealing with database seeding in Laravel, especially when constraints are involved, it can indeed be tricky to debug. Here are some steps and tips to help you manage and debug your seeders effectively:

Order of Seeding: Ensure that you are seeding your tables in the correct order, respecting the foreign key constraints. For example, if you have a users table and a posts table where posts has a foreign key referencing users, you should seed users first.
Disable Foreign Key Checks: Temporarily disable foreign key checks during seeding to avoid constraint violations. You can do this by adding the following lines in your seeder:
```
use Illuminate\Support\Facades\DB;

DB::statement('SET FOREIGN_KEY_CHECKS=0;');
// Your seeding logic here
DB::statement('SET FOREIGN_KEY_CHECKS=1;');
```
Use dd() for Debugging: If you want to use dd() to debug your seeder and see the output in the browser, you can run your seeder through a route or a controller method temporarily. This is not recommended for production but can be useful for debugging:
```
Route::get('/run-seeder', function () {
    Artisan::call('db:seed', [
        '--class' => 'YourSeederClassName'
    ]);
    dd('Seeding completed');
});
```
Accessing /run-seeder in your browser will execute the seeder and display the dd() output.
Log Output: Instead of using dd(), consider logging the output to a file. This can be done using Laravel's logging facilities:
```
use Illuminate\Support\Facades\Log;

Log::info('Seeding started');
// Your seeding logic here
Log::info('Seeding completed');
```
Check the storage/logs/laravel.log file for the output.
Use Transactions: Wrap your seeding logic in a transaction to ensure that if something goes wrong, the database state is not left in an inconsistent state:
```
DB::transaction(function () {
    // Your seeding logic here
});
```

By following these steps, you should be able to manage and debug your database seeding process more effectively in Laravel.

Tray2

1 year ago

Level 73

I mainly use seeders for my tests, and for data that is supplied with the application, such as categories, formats, and that kind of stuff. Basically data that isn't supplied by the user.

For tests I mostly use a factory class for the model, but sometimes a seeder may be a better option.

When it comes to migrating data from one database to another, or migrating a application from regular php and an old db model to Laravel and it's conventions I would not create seeders for it, I would script it step by step.

Seeding with constraints is always a bit tricky, but you can use the model factory to set that up if needed, but my recommendation is not tu use seeders for that.

Merklin

1 year ago

Level 7

Even with constrains, you can send predefined data to the factory when creating records. Lets say you have a Project class with 2 fields: name and user, where the column for the user is named user_id and is constrained to the users table. A simple factory will be like this:

public function definition(): array
    {
            'name' => $this->faker->name(),
            'user_id' => User::all()->random(1)->first(),
    }

In fact, you can completely omit the user_id in the factory, and when calling it from a seeder or test use:

Project::factory(5)->create(['user_id' => your_logic_here])

This way, you can even call the factory within a loop:

foreach(User:all() as $user) {
    Project::factory(5)->create(['user_id' => $user->id])
}

martinbean

1 year ago

Level 80

@dmjcv91913 I’m struggling to understand what problem you’re trying to describe? If you have a database schema with foreign key constraints then you need to seed records in a logical order (i.e. creating dependant records first). You’d have to do that regardless if you were using an ORM or not.

1 like

dmjcv91913

1 year ago

Level 1

I am learning a lot about the ORM in this process. I am definitely mapping out the logic step by step and the database constraint mapping. it is a little different than traditional relational databases with constraints, but most is fairly similar. I will say there are some tricks to using factories and seeders when dealing with relationship constraints and everyone should painstakingly go through different ways to logically code them. I traditionally just run code in PHP and inject the data into MySQL. I still may go that route, but i will say it is very handy to have these tools available in the platform. Thank you all for your responses. I will post any interesting code I come up with relative to the syntax and modeling requirements of the project i am working on in the insurance space.

martinbean

1 year ago

Level 80

@dmjcv91913 I’m still struggling to understand just what it is you’re talking/asking about?

If you’re seeding a database, then you need to create dependent records first. That’s just the way things are. You can’t create “child” records without creating their parents first. This is the case no matter what language or paradigm or framework you use.

Please or to participate in this conversation.