Include Auto Increment in Schema Dump

[liamduckett]

Hey,

I recently dumped my MySQL schema - and was surprised to find that my primary key starting value starting value was not included in the schema dump.

public function up(): void
{
    Schema::create('users', function (Blueprint $table) {
        $table->id()->startingValue(123);
    });
}

Results in:

CREATE TABLE `users` (
  `id` bigint unsigned NOT NULL AUTO_INCREMENT,
  PRIMARY KEY (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci;

Whilst I would have expected

CREATE TABLE `users` (
  `id` bigint unsigned NOT NULL AUTO_INCREMENT,
  PRIMARY KEY (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci AUTO_INCREMENT=100;

Upon taking a deeper look at this, I found the following method:

// src/Illuminate/Database/Schema/MySqlSchemaState.php (34)

/**
 * Remove the auto-incrementing state from the given schema dump.
 *
 * @param  string  $path
 * @return void
 */
protected function removeAutoIncrementingState(string $path)
{
    $this->files->put($path, preg_replace(
        '/\s+AUTO_INCREMENT=[0-9]+/iu',
        '',
        $this->files->get($path)
    ));
}

...which is why I'm reporting it as a feature request, rather than a bug report (which I had originally planned) - it seems intentional.

Can anyone shed some light on the reason for this? The best I could come up with, was for consistency with Postgres (which is the only other RDBMS to support setting a startingValue). For reference, a Postgres schema dump doesn't include auto increment starting values, but also doesn't feature any code to specifically do it.

[henzeb]

Why do you need auto increment?

[henzeb]

Yes

[liamduckett]

Truthfully I don't know - it was a project I inherited. I'm not too sure it'd be relevant though?

[henzeb]

Oh it's relevant alright. They want the code to be slim as possible and even if it takes barely ane extra code, Taylor is dismissing anything he doesn't think it's likely to be used by many. As the feature usually is used for setting up dev environments and tests, I doubt a chance of this change s getting in with a solid use case.

Remember auto increment changes automatically if a record with manual ID comes in with a higher number than the current last increment..

[liamduckett]

That's fair. I can't speak for how often startingValue is used (I can't see many use cases), but I do think schema dump in general is quite likely to get an increased usage in the near future. The Laravel 10 -> Laravel 11 upgrade suggests it, in the case of projects that have "changed" columns in migrations: https://laravel.com/docs/11.x/upgrade#modifying-columns.

Any project that uses startingValue following this guide, will likely end up with unexpected (and not immediately obvious) changes in behaviour.

[henzeb]

I Don't think that guide is meant for production. Maybe describe your use case then?