pseudo code is as follows

// b 
$index = 0
// b  
$contains = []
// for  a
for ():
    if (a [i]>= 500) 
        continue
    else 
        a[i].num  = b[$index].num
        $contains.push(b[$index].id)
        $indexPP

Sorry, php learned but forgot that if b can fill an at will, it's OK
if b has to be greater than 500 to fill, you can just get rid of b less than 500

Let me also say a thought:

1. use array_filter to screen out the conditions in the array a
2. for the first time array_replace , replace the sifted value
in [1] with the array b.
3. for the second time array_replace , replace the value in array a with [2]
4. run through the loop, using array_key_exists to determine whether there is a key, named 'id'' for each array element in array a. If so, put the whole push into a new array and unset drop
.

may not be very elegant. It's just one more angle.

the following is key in $a is id

 $a = collect($a)->map(function($a1, $key) use($b){
        if($a1['num'] < 500){
            $b1 = collect($b)->reject(function($b1) use($key){
               return $b1['id'] != $key;
            });
            $b1 = $b1->flatMap(function($xx){
                return $xx;
            });

            if (!empty($b1->toArray())){
                $a1['num'] = $b1['num'];
                $a1['index'] = $b1['id'];
                return $a1;
            }  
        }
    })->filter(function($v){
        return !empty($v);
    });

    dd($a);