从mongo DB集合中的对象数组中查询唯一文档

Question

我需要一个查询从"A“集合中删除重复的文档。一旦重复被删除，那么我将创建一个"A“集合的索引。

"A“藏书有近132920份文件。

请查找"A“集合Mongo DB文档结构，

/*1 */
    {
        "_id" : ObjectId("5d654ed25616a9d461bc0aa7"),
        "aC" : "AB",
        "aI" : "ABCD",
        "aN" : "000000002001",
        "aT" : "CA",
        "bID" : NumberLong(0),
        "pD" : "2019-06-18",
        "transactions" : [ 
            {
                "seqNo" : 1,
                "tC" : "S"
            }, 
            {
                "seqNo" : 2,
                "tC" : "S"
            },
            {
                "seqNo" : 2,
                "tC" : "S"
            },
            {
                "seqNo" : 1,
                "tC" : "S"
            },
            {
                "seqNo" : 3,
                "tC" : "S"
            }
        ]
    }

/*2*/
{
        "_id" : ObjectId("5d654ed25616a9d461bc0aa8"),
        "aC" : "AB",
        "aI" : "ABCD",
        "aN" : "000000002002",
        "aT" : "CA",
        "bID" : NumberLong(0),
        "pD" : "2019-06-18",
        "transactions" : [ 
            {
                "seqNo" : 1,
                "tC" : "S"
            }, 
            {
                "seqNo" : 2,
                "tC" : "S"
            },
            {
                "seqNo" : 2,
                "tC" : "S"
            },
            {
                "seqNo" : 1,
                "tC" : "S"
            },
            {
                "seqNo" : 3,
                "tC" : "S"
            }
        ]
    }

预期产出：

/*1*/
{
        "_id" : ObjectId("5d654ed25616a9d461bc0aa7"),
        "aC" : "AB",
        "aI" : "ABCD",
        "aN" : "000000002001",
        "aT" : "CA",
        "bID" : NumberLong(0),
        "pD" : "2019-06-18",
        "transactions" : [ 
            {
                "seqNo" : 1,
                "tC" : "S"
            }, 
            {
                "seqNo" : 2,
                "tC" : "S"
            },
            {
                "seqNo" : 3,
                "tC" : "S"
            }
        ]
    }

/*2*/
{
        "_id" : ObjectId("5d654ed25616a9d461bc0aa8"),
        "aC" : "AB",
        "aI" : "ABCD",
        "aN" : "000000002002",
        "aT" : "CA",
        "bID" : NumberLong(0),
        "pD" : "2019-06-18",
        "transactions" : [ 
            {
                "seqNo" : 1,
                "tC" : "S"
            }, 
            {
                "seqNo" : 2,
                "tC" : "S"
            },
            {
                "seqNo" : 3,
                "tC" : "S"
            }
        ]
    }

在从“事务数组”中删除重复项之后。我将创建"A“集合的主要索引如下，

{
"aC" : 1.0,
"aI" : 1.0,
"aN" : 1.0,
"aT" : 1.0,
"bID" : 1.0,
"pD" : 1.0

}

请帮助我从"A“集合中移除重复的文档，以便我可以启用"A”集合的索引吗？

在查询方面需要帮助。

提前谢谢。这会很有帮助的。

Answer 1

有一种方法你可以做到。

db.collection.update({
  "transactions": {"$type": "array"}
},
[
  {
    "$set": {
      "transactions": {"$setUnion": "$transactions"}
    }
  }
],
{"multi": true}
)

在mongoplayground.net上试一试。

编辑

我不知道是mongoplayground.net还是MongoDB的“奇怪”/bug，但我怀疑有什么东西“重新排序”了"transactions"数组对象中的字段。

这里有一个“解决方案”，它在"$setUnion"之前重写每个元素。

db.collection.update({
  "transactions": {"$type": "array"}
},
[
  {
    "$set": {
      "transactions": {
        "$setUnion": {
          "$map": {
            "input": "$transactions",
            "as": "xaction",
            "in": {
              "seqNo": "$$xaction.seqNo",
              "tC": "$$xaction.tC"
            }
          }
        }
      }
    }
  }
],
{"multi": true}
)

在mongoplayground.net上试一试。