如何在php中对mongodb文档进行分组

Question

I am using mongo server 3.6 and php as backend language. Also I am using the latest php-mongo library to talk to new updated drivers.

I have a collection with 10 million records in it like so :

[
  {
    "did": "123456",
    "did_usage": "1",
    "did_timestamp": "15012"
  },
  {
    "did": "4567811",
    "did_usage": "1",
    "did_timestamp": "15013"
  },
  {
    "did": "46465464",
    "did_usage": "2",
    "did_timestamp": "15014"
  },
  {
    "did": "7894446",
    "did_usage": "2",
    "did_timestamp": "15015"
  },
  {
    "did": "65646131",
    "did_usage": "3",
    "did_timestamp": "15016"
  },
  {
    "did": "7989464",
    "did_usage": "2",
    "did_timestamp": "15017"
  },
  {
    "did": "651651664",
    "did_usage": "1",
    "did_timestamp": "15018"
  }.......
]

Now I want to find a single unique document which have least did usage and least did timestamp.

Till now I have reached to find the single unique document using :

$sample = array('$sample' => array('size' => 1));
$pipeline = array($match, $group, $project, $sample);
$cursor = $collection->aggregate($pipeline);

I want some help regarding the $group . I have tried this

$group = array('$group' => array('_id' => '$did_usage', 'did_usage_timestamp' => array('$min' => '$did_usage_timestamp')));

But this is not working as expected.

Answer 1

Even without knowing what $match and $project are doing in your original code, we can assume that $group is operating on a subset of the collection data since it is only preceded by $match in the pipeline. Based on only the example collection data and the $group stage, it's evident that the $dig_usage_timestamp field path used with the $min operator references a non-existent field in the documents entering the $group stage.

When testing this locally, did_usage_timestamp in the randomly selected output document is null :

<?php

require 'vendor/autoload.php';

$client = new MongoDB\Client;
$collection = $client->test->foo;

$collection->drop();
$collection->insertMany([
    ["did" => "123456", "did_usage" => "1", "did_timestamp" => "15012"],
    ["did" => "4567811", "did_usage" => "1", "did_timestamp" => "15013"],
    ["did" => "46465464", "did_usage" => "2", "did_timestamp" => "15014"],
    ["did" => "7894446", "did_usage" => "2", "did_timestamp" => "15015"],
    ["did" => "65646131", "did_usage" => "3", "did_timestamp" => "15016"],
    ["did" => "7989464", "did_usage" => "2", "did_timestamp" => "15017"],
]);

$cursor = $collection->aggregate([
    ['$group' => ['_id' => '$did_usage', 'did_timestamp' => ['$min' => '$did_usage_timestamp']]],
    ['$sample' => ['size' => 1]],
]);

var_dump($cursor->toArray());

This outputs something similar to:

array(1) {
  [0]=>
  object(MongoDB\Model\BSONDocument)#14 (1) {
    ["storage":"ArrayObject":private]=>
    array(2) {
      ["_id"]=>
      string(1) "1"
      ["did_timestamp"]=>
      NULL
    }
  }
}

Changing the $min operator's field path to $did_timestamp should resolve the issue.