Home / Questions / 429907
Accepted
elapez
Asked: 2022-02-21 19:15:57 +0800 CST 2022-02-21 19:15:57 +0800 CST

Process QueryBuilder chunks with chunk()

  • 772

I have a method of a class that I extend from ELOQUENT that has to process very large collections of the order of more than 800,000 items and I have realized that it eventually causes some difficulties in the service, so I decided according to the advice of the documentation of Laravel, use chunk() method from QueryBuilder. The process is to read the existing entries in a certain date range and then calculate averages in time subdivisions of the entire collection, to display in a behavior graph. The problem is that the chunk() method perfectly processes each chunk calculating the averages of each input segment but I can't get it to return those averages, I can print them from an "echo" but as by default it does a loop iterating between each chunk of the collection, at the end it only returns a "

public static function gghh($campo, $marca = [], $corte=5000)
{
    $ini = isset($marca["ini"]) ? $marca["ini"] : Carbon::now()->subDay()->format('Y-m-d');
    $fin = isset($marca["fin"]) ? $marca["fin"] : Carbon::now()->format('Y-m-d');

    return self::whereBetween("fecha", [$ini, $fin])->chunk($corte,function ($regs) use ($campo){

        //echo $regs->avg($campo)."<br />";
        //echo $regs->first()->fecha."<br />";
    
        if(!isset($trozos))
        {
           $trozos = [];
        }
        
        array_push($trozos, ["dato" => $regs->avg($campo), "fecha" => $regs->first()->fecha]);

        return $trozos
    });
}

I know that the process is done correctly because if I remove the comments from the two "echoes", the values ​​are displayed correctly, but I need these same values ​​to be returned to me for a later process. As it is now shown, the method only returns a boolean value (TRUE). Any ideas to obtain something like the Array that follows?

[
  ["dato"=>145.58, "fecha"=>"2020-08-12"],
  ["dato"=>128.89, "fecha"=>"2020-09-12"],
  ["dato"=>132.14, "fecha"=>"2020-10-12"],
  ["dato"=>136.62, "fecha"=>"2020-11-12"],
  ...
]
laravel

1 Answers

  1. Best Answer

    It's a scope issue. If $trozosit doesn't exist and you declare it inside the loop, it doesn't exist outside of the loop, nor in the next loop. The method chunkis not one mapthat you can return from, but only one eachin easy installments.

    The only effect it has return $trozosis that if in any iteration this variable is falsy it is assumed that you don't want to process the batches any further.

    I would try:

      $trozos = [];

  return self::whereBetween("fecha", [$ini, $fin])
      ->chunk($corte, function ($regs) use ($campo, &$trozos){
      
       $trozos[] = [
           "dato" => $regs->avg($campo), 
           "fecha" => $regs->first()->fecha
       ]);
  });

    Note that I pass it $trozosby reference, since if I pass it as a value, it would be expected not to affect its value in the upper scope.

    • 2