使用save()方法保存新的Model时返回关系数据和新的Model object

Question

$user = User::find(1);
$post = new Post();
$user->posts()->save($post);

I want to access the new Post 's User without making an additional query to the database. It queries the database for the User if I do the following:

$userPost = $post->user;

I'm aware I could just do $userPost = $user , but for certain reasons, I don't want to do it that way.

So my question is, would it be possible to eager load the User when saving the new Post with save() ?

Answer 1

I just try the following way to save new post with minimal query.

$user = User::with('posts')->find(1);

// create new post , but not save it yet
$post = new Post(['title' => 'lorem ipsum', 'body' => 'something something']);

// push it to user posts collection
$user->posts->push($post);

// and now save it
$user->posts()->save($post);

// then check the content, post should now have id and timestamps
dd($user, $user->posts);

The new Post is there with id and timestamps, without need to reload all posts, saving you a bit performance and execution time.

One might asked, how many queries were executed?

1. "select * from "users" where "users"."id" =? limit 1"
2. "select * from "posts" where "posts"."user_id" in (1)"
3. "insert into "posts" ("title", "body", "user_id", "updated_at", "created_at") values (?, ?, ?, ?, ?) returning "id""

I used DB::enableQueryLog() before transactions and DB::getQueryLog() after transactions to track these queries.

...

And regarding your question: $post->user .
Unfortunately this will still trigger an extra query .

"select * from "users" where "users"."id" = 1 limit 1"

---

Now lets compare to good old add post to user steps. (2nd option)

• $id = 1;
• $user = User::with('posts')->find($id); // needs 2 queries
• $post = Post::create(['user_id' => $id, 'title'.... ]); // 3rd query
• $user->load('posts'); // 4th query (but this is optional)
• $userPost = $post->user; // and last query

In total will need to execute 4 or 5 queries, depend on do you need to load all posts as user's children.

In term of performance, cpu usage and execution time are almost the same, with the 1st steps above is a bit faster in execution time.

So at the end it is back to you.
What do you really need in the system you are currently building?
Keep the $user->posts with new post added, that choose 1st.
But if you just want directly to catch $userPost, than use 2nd option.

Note: no need to use $user = User::find($id) if you already know that user $id is exists.