我使用mongo server 3.6和php作为后端语言。 此外,我正在使用最新的php-mongo库与新的更新驱动程序进行对话。
我有一个包含1000万条记录的集合,如下所示:
[
{
"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"
}.......
]
现在我想找到一个唯一一个使用率最少且时间戳最少的独特文档。
到目前为止,我已经使用以下方法找到了唯一的独特文档:
$sample = array('$sample' => array('size' => 1));
$pipeline = array($match, $group, $project, $sample);
$cursor = $collection->aggregate($pipeline);
我想了解$group
的一些帮助。
我试过这个
$group = array('$group' => array('_id' => '$did_usage', 'did_usage_timestamp' => array('$min' => '$did_usage_timestamp')));
但这并没有像预期的那样发挥作用。
答案 0 :(得分:2)
即使不知道原始代码中$match
和$project
正在做什么,我们也可以假设$group
正在对集合数据的子集进行操作,因为它前面只有{ {1}}在管道中。仅基于示例集合数据和$match
阶段,显然与$group
运算符一起使用的$dig_usage_timestamp
field path引用了进入该文档的文档中不存在的字段$min
阶段。
在本地测试时,随机选择的输出文档中的$group
为did_usage_timestamp
:
null
这会输出类似于:
的内容<?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($cursor->toArray());
将array(1) {
[0]=>
object(MongoDB\Model\BSONDocument)#14 (1) {
["storage":"ArrayObject":private]=>
array(2) {
["_id"]=>
string(1) "1"
["did_timestamp"]=>
NULL
}
}
}
运营商的字段路径更改为$min
可以解决问题。