我正在尝试找到一个 mongoDB 脚本，该脚本将查看一个集合，其中存在同一文档的多个记录，并且只向我提供每个文档的最新版本作为结果集。

我无法用英语比上面更好地解释它，但也许下面的这个小 SQL 可以进一步解释它。我希望每个文档都按 transaction_reference 但只有最新版本 (object_creation_date)。

select 
    t.transaction_reference, 
    t.transaction_date, 
    t.object_creation_date,
    t.transaction_sale_value
from MyTable t
inner join (
    select 
        transaction_reference, 
        max(object_creation_date) as MaxDate
    from MyTable
    group by transaction_reference
) tm 
    on t.transaction_reference = tm.transaction_reference 
    and t.object_creation_date = tm.MaxDat

之所以有同一个文档的多个版本，是因为我想存储一个事务的每个迭代。我第一次收到文件时，它可能处于 UNPAID 的 transaction_status 然后我再次收到相同的交易，这次 transaction_status 是 PAID。

一些分析将对所有唯一交易求和，而另一些分析可能是测量状态为 UNPAID 的文档与下一个 PAID 文档之间的时间距离。

根据要求，这里有两个文件:

{
"_id": {
    "$oid": "579aa337f36d2808839a05e8"
},
"object_class": "Goods & Services Transaction",
"object_category": "Revenue",
"object_type": "Transaction",
"object_origin": "Sage One",
"object_origin_category": "Bookkeeping",
"object_creation_date": "2016-07-05T00:00:00.201Z",
"party_uuid": "dfa1e80a-5521-11e6-beb8-9e71128cae77",
"connection_uuid": "b945bd7c-7988-4d2a-92f5-8b50ab218e00",
"transaction_reference": "SI-1",
"transaction_status": "UNPAID",
"transaction_date": "2016-06-16T00:00:00.201Z",
"transaction_due_date": "2016-07-15T00:00:00.201Z",
"transaction_currency": "GBP",
"goods_and_services": [
    {
        "item_identifier": "PROD01",
        "item_name": "Product One",
        "item_quantity": 1,
        "item_gross_unit_sale_value": 1800,
        "item_revenue_category": "Sales Revenue",
        "item_net_unit_cost_value": null,
        "item_net_unit_sale_value": 1500,
        "item_unit_tax_value": 300,
        "item_net_total_sale_value": 1500,
        "item_gross_total_sale_value": 1800,
        "item_tax_value": 300
    }
],
"transaction_gross_value": 1800,
"transaction_gross_curr_value": 1800,
"transaction_net_value": 1500,
"transaction_cost_value": null,
"transaction_payments_value": null,
"transaction_payment_extras_value": null,
"transaction_tax_value": 300,
"party": {
    "customer": {
        "customer_identifier": "11",
        "customer_name": "KP"
    }
}
}

和现在付费的第二个版本

{
"_id": {
    "$oid": "579aa387f36d2808839a05ee"
},
"object_class": "Goods & Services Transaction",
"object_category": "Revenue",
"object_type": "Transaction",
"object_origin": "Sage One",
"object_origin_category": "Bookkeeping",
"object_creation_date": "2016-07-16T00:00:00.201Z",
"party_uuid": "dfa1e80a-5521-11e6-beb8-9e71128cae77",
"connection_uuid": "b945bd7c-7988-4d2a-92f5-8b50ab218e00",
"transaction_reference": "SI-1",
"transaction_status": "PAID",
"transaction_date": "2016-06-16T00:00:00.201Z",
"transaction_due_date": "2016-07-15T00:00:00.201Z",
"transaction_currency": "GBP",
"goods_and_services": [
    {
        "item_identifier": "PROD01",
        "item_name": "Product One",
        "item_quantity": 1,
        "item_gross_unit_sale_value": 1800,
        "item_revenue_category": "Sales Revenue",
        "item_net_unit_cost_value": null,
        "item_net_unit_sale_value": 1500,
        "item_unit_tax_value": 300,
        "item_net_total_sale_value": 1500,
        "item_gross_total_sale_value": 1800,
        "item_tax_value": 300
    }
],
"transaction_gross_value": 1800,
"transaction_gross_curr_value": 1800,
"transaction_net_value": 1500,
"transaction_cost_value": null,
"transaction_payments_value": null,
"transaction_payment_extras_value": null,
"transaction_tax_value": 300,
"party": {
    "customer": {
        "customer_identifier": "11",
        "customer_name": "KP"
    }
}
}

谢谢你的支持，马特

最佳答案

如果我正确理解了这个问题，你可以使用这样的东西

db.getCollection('yourTransactionsCollection').aggregate([
    {
        $sort: {
            "transaction_reference": 1,
            "object_creation_date": -1
        }
    },
    {
        $group: {
            _id: "$transaction_reference",
            "transaction_date": { $first: "$transaction_date" },
            "object_creation_date": { $first: "$transaction_date" },
            "transaction_sale_value": { $first: "$transaction_sale_value" }
        }
    }
])

输出如下结果

{
    "_id" : "SI-1",
    "transaction_date" : "2016-06-16T00:00:00.201Z",
    "object_creation_date" : "2016-06-16T00:00:00.201Z",
    "transaction_sale_value" : null
}

请注意，您可以将 $sort 更改为仅包含 object_creation_date，但我同时包含了 transaction_reference 和 object_creation_date 因为我认为为它们创建一个复合索引而不只是创建日期是有意义的。根据您的索引进行调整，使 $sort 达到一个。
此外，没有文档字段 transaction_sale_value，因此结果中它的 null。也许您错过了它，或者它只是不在您的示例文档中，但我认为您已经理解并可以根据您的需要进行调整。

关于mongodb - 在mongoDB中按日期查询文档的最新版本，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/38650002/