Best Practice? Using 'auth()->user()' or '$user = auth()->user()'?

use the $user variable if you are going to call auth()->user()** more than once.

why ? .. think about it.

lets say that in a function you have a call to atuh()->user()->* (where * is whatever) ... well ... how PHP will resolve that??

It will go step by step (-> by ->) ... so ... now imagine you have more than one call to auth()->user()->** , well , PHP will have to resolve those expressions over and over again, .......... if you are looking for performance ... use the $user variable containing auth()->user() result inside it.

Either way it's really the same thing to php. If you assign it to $user, then $user is just pointing to the same memory address as auth()->user(). It's not making a separate copy or something. If you do

$user = auth()->user();
echo $user->name; // 'Joe' or something

$user->name = 'changed it';

echo $user->name; // 'changed it'
echo auth()->user()->name; // 'changed it'

So they are the same thing. I do $user = auth()->user() all the time (as well as on other things besides the user object), to keep code more clear and reduce it. If I'm just using it once in a controller or something, I just directly use auth()->user(). If I'm referencing it multiple times, I assign it to a variable.

Hey @cronix - you got a face!

@Snapey I didn't want to deprive you guys any longer. I know it was tough and appreciate everyone's patience lol.

Is the same result, but I have my doubts if its "the same for php" unless php has some cache inside it.

please correct me if I am wrong:

Whe PHP resolves this expression "auth()->user()" , basically what happens :

1. resolves auth()
2. take the result in 1) and execute the "user()" method on it
3. return result obteined in 2)

so if I have in my code "auth()->user()" over and over again I would have the 3 steps over and over again.

But if we assign $user = auth()->user() , to a variable, you already have those 3 steps done.

That is why even though you get the same result, ... is not the same for PHP in terms of performance execution.

please, correct me if I am wrong, interesting subject.

Is the same result, but I have my doubts if its "the same for php" unless php has some cache inside it.

For that we have Bytecode Cache's, right? :)

@Helmchen , probably, if someone can provide info related, I would appreciate

Of the many 10's of thousands of lines of code that are run, I doubt it makes a single jot of difference.

Do whatever is easiest to type and understand

I'm not a scientist but

$start = microtime(true);

$user = Auth::user();
for($i=0; $i < 1000000; $i++)
{
    $id = $user->id;
}

// 2.1193430423737
$first = microtime(true) - $start;

$start = microtime(true);

for($i=0; $i < 1000000; $i++)
{
    $id = Auth::user()->id;
}

// 8.5643661022186
$second = microtime(true) - $start;

it actually makes a difference, but i doubt you will call that 1.000.000 times.

If you do this 5 or 10 times, the result is a unit for which I do not even know the name, definitely negligible.

So practically, @Cronix is right, technically it makes a difference, though not a big one.

@Helmchen thanks for those numbers

on an average site I could see this being negligible as you of course wouldn't use anywhere near 1 000 000 calls.

However if you had a app that started to get popular (for example Twitter, Facebook, etc) I could see this making an impact with methods using user info being called millions of times.

I think @Cronix response covers best practise. If only 1 call use auth()->user()->... more then 1 call use $user = auth()->user().

@jcmargentina if you make "$user = auth()->user()" and then change "$user->name", without saving it, then echo "auth()->user()->name" and it also has updated name wouldn't this have to indicate that it is saved in cache and $user is a pointer to the same data as auth()->user()?? $user still being faster as it is a direct path to the same data... However this is just my thought process based off Cronix's reply.

However if you had a app that started to get popular (for example Twitter, Facebook, etc) I could see this making an impact with methods using user info being called millions of times.

We're talking about a single request cycle though to generate the output for a single page. I certainly hope no one is calling it a million times on the same page

We're talking about a single request cycle though to generate the output for a single page. I certainly hope no one is calling it a million times on the same page

I meant that a single request cycle could be called millions of times (since there are millions of users). I was meaning that if user data is accessed multiple times each request cycle then it would start to become a factor (albeit small) with so many individual "single request cycles".

My takeaway is that it doesn't really matter but best practise would be to assign it to a variable if used more then once or twice.

No single server in the world will handle 1,000,000 concurrent users, but still

My takeaway is that it doesn't really matter but best practise would be to assign it to a variable if used more then once or twice.

I think you can leave it that way :)

I love this forum.