mongodb - 我怎样才能从数组中删除重复的项目(复杂对象)

标签 mongodb aggregation-framework

在每个文档中,

records 是一个包含许多重复对象的数组。

buy_items 中也包含许多重复的项目。

如何清理重复项?

原始文档:

{
  "_id": "0005d116qwwewdq82a1b84f148fa6027d429f3e",
  "records": [
    {
      "DATE": new Date("1996-02-08T08:00:00+0800"),
      "buy_items": [
        "5210 ",
        "5210 ",
        "5210 "
      ]
    },
    {
      "DATE": new Date("1996-02-08T08:00:00+0800"),
      "buy_items": [
        "5210 ",
        "5210 ",
        "5210 "
      ]
    }
    {
      "DATE": new Date("2012-12-08T08:00:00+0800"),
      "buy_items": [
        "5210 ",
        "1234 ",
        " "
      ]
    }        
    ]
}

预期输出:

{
  "_id": "0005d116qwwewdq82a1b84f148fa6027d429f3e",
  "records": [
    {
      "DATE": new Date("1996-02-08T08:00:00+0800"),
      "buy_items": [
        "5210 "
      ]
    },
    {
      "DATE": new Date("2012-12-08T08:00:00+0800"),
      "buy_items": [
        "5210 ",
        "1234 ",
        " "
      ]
    }    
    ]
}

使用 Michaels 解决方案,输出可能如下所示

{
  "_id": "0005d116qwwewdq82a1b84f148fa6027d429f3e",
  "records": [
    "date": new Date("1996-02-08T08:00:00+0800"),
      "buy_items": [
        "5210 "
        "1234 ",
        " "
      ]
    ]
}

最佳答案

您可以使用 aggregation framework 删除重复的对象

db.collection.aggregate(
    [
        { $unwind: "$records" }, 
        { $unwind: "$records.buy_items" }, 
        { $group: { "_id": {id: "$_id", date: "$records.DATE" }, buy_items: { $addToSet: "$records.buy_items" }}}, 
        { $group: {"_id": "$_id.id", records: { $push: {"date": "$_id.date", "buy_items": "$buy_items" }}}}, { $sort: { "records.0.date": 1 }} ,
        { $out: "collection" }
    ]
)

$out运算符可让您将聚合结果写入指定集合或替换现有集合。


使用 "Bulk" 效果更好操作

var bulk = bulk = db.collection.initializeOrderedBulkOp(),
    count = 0;

db.collection.aggregate([
    { "$unwind": "$records" }, 
    { "$project": { 
        "date": "$records.DATE", 
        "buy_items": { "$setIntersection": "$records.buy_items" }
    }}, 
    { "$unwind": "$buy_items" }, 
    { "$group": { 
        "_id": { "id": "$_id", "date": "$date" }, 
        "buy_items": { "$addToSet": "$buy_items" }
    }},
    { "$group": { 
        "_id": "$_id.id", 
        "records": { "$push": { 
            "date": "$_id.date", 
            "buy_items": "$buy_items" 
        }}
    }}
]).forEach(function(doc) { 
       bulk.find({"_id": doc._id}).updateOne({
       "$set": { "records": doc.records }
       }); 
       count++; 
       if (count % 500 == 0) {   
           bulk.execute();    
           bulk = db.collection.initializeOrderedBulkOp(); 
       } 
})

if (count % 500 != 0)
    bulk.execute();

结果:

{
    "_id" : "0005d116qwwewdq82a1b84f148fa6027d429f3e",
    "records" : [
            {
                    "date" : ISODate("2012-12-08T00:00:00Z"),
                    "buy_items" : [
                            " ",
                            "1234 ",
                            "5210 "
                    ]
            },
            {
                    "date" : ISODate("1996-02-08T00:00:00Z"),
                    "buy_items" : [
                            "5210 "
                    ]
            }
    ]
}

关于mongodb - 我怎样才能从数组中删除重复的项目(复杂对象),我们在Stack Overflow上找到一个类似的问题: https://stackoverflow.com/questions/30613568/

相关文章:

node.js - 如何在Keystonejs后端显示上传的图片

node.js - Array.push 似乎不适用于 MongoDB

尝试使用 MongoDB BulkWriteOperation 时出现 java.lang.IllegalStateException

java - 使用 Java 更新 MongoDB 中的特定字段而不是整个文档

python - $graphLookup 管道阶段后过滤结果

mongodb - 如何在 MongoDB 中使用数组位置 $ 和聚合投影

mongodb - 在 MongoDB 上使用 UUID 而不是其字符串表示作为 _id 的性能提升?

MongoDB 通过匹配子文档按相关性排序

mongodb - 如何在分组时使用 $cond 检查同一数组中的空值进行聚合操作

mongodb - 使用带有条件的 $graphLookup