在 Laravel 測(cè)試中管理數(shù)據(jù)庫(kù)狀態(tài)的方法包括使用 RefreshDatabase、選擇性播種數(shù)據(jù)、謹(jǐn)慎使用事務(wù)和必要時(shí)手動(dòng)清理。1. 使用 RefreshDatabase trait 自動(dòng)遷移數(shù)據(jù)庫(kù)結(jié)構(gòu)，確保每次測(cè)試都基于干凈的數(shù)據(jù)庫(kù)；2. 通過(guò)調(diào)用特定種子填充必要數(shù)據(jù)，結(jié)合模型工廠生成動(dòng)態(tài)數(shù)據(jù)；3. 使用 DatabaseTransactions trait 回滾測(cè)試更改，但需注意其局限性；4. 在無(wú)法自動(dòng)清理時(shí)，手動(dòng)截?cái)啾砘蛑匦虏シN數(shù)據(jù)庫(kù)。這些方法根據(jù)測(cè)試類型和環(huán)境靈活選用，以保證測(cè)試的可靠性和效率。

Testing in Laravel often involves working with a real database, and managing the state of that database between tests is crucial for reliable and predictable outcomes. The goal is to make sure each test runs in isolation, without side effects from previous tests or leftover data.

Here are some practical ways to manage your database state when testing in Laravel.

Use Migrations with RefreshDatabase

One of the most common and effective strategies in Laravel is using the RefreshDatabase trait. It automatically migrates your database schema before your tests run, ensuring a clean slate every time.

This works especially well with SQLite in-memory databases or MySQL/PostgreSQL during testing. Just add the trait to your test class:

use Illuminate\Foundation\Testing\RefreshDatabase;

class ExampleTest extends TestCase
{
    use RefreshDatabase;

    // Your tests here
}

It’s fast and keeps your database structure up to date without manually truncating tables or re-seeding everything each time.

Seed Data Selectively

Sometimes you need a certain set of data to exist before running your tests — this is where seeding comes in handy. You can call specific seeders inside your tests to populate the necessary data.

For example:

$this->seed(UserTableSeeder::class);

Avoid seeding the entire database unless absolutely necessary. Instead, only seed what your test requires. This keeps tests faster and more focused.

You can also combine seeding with model factories to generate dynamic test data on the fly:

User::factory()->count(5)->create();

This way, you get consistent but realistic data without overloading the database.

Use Transactions (with Caution)

Another approach is wrapping each test in a database transaction and rolling it back afterward. Laravel supports this via the DatabaseTransactions trait.

use Illuminate\Foundation\Testing\DatabaseTransactions;

class ExampleTest extends TestCase
{
    use DatabaseTransactions;

    // Tests go here
}

This method is fast because it doesn’t truncate or refresh the database — it just rolls back any changes after the test finishes. However, be careful: if your code uses Artisan commands or interacts with the DB outside Eloquent (like raw queries), transactions might not roll back properly.

Also, this doesn't work well with drivers like SQLite when using an in-memory database, since transactions behave differently.

Clean Up After Yourself (When Needed)

In some cases — especially feature or browser tests — you may end up with data that isn’t cleaned up by migrations or transactions. In these situations, consider manually truncating tables at the start or end of a test.

You can do this with:

Artisan::call('db:seed', ['--class' => 'DatabaseSeeder', '--force' => true]);

Or truncate specific tables:

DB::table('users')->truncate();

Just keep in mind that truncating is slower than using transactions or migrations, so use it sparingly.

Managing database state in Laravel tests doesn’t have to be complicated. Pick the right tool based on your test type and environment — usually RefreshDatabase covers most needs. Combine it with selective seeding and model factories for best results. Transactions can speed things up but come with caveats. And sometimes, a manual cleanup is the only way to go.

基本上就這些。

以上是管理數(shù)據(jù)庫(kù)狀態(tài)進(jìn)行Laravel測(cè)試的詳細(xì)內(nèi)容。更多信息請(qǐng)關(guān)注PHP中文網(wǎng)其他相關(guān)文章！