c# - 条件聚合器的 MongoDb.Linq 等效查询

Question

我按以下方式使用 MongoDb.Linq 对我的数据进行分组:

group r1 by new { r1.SupId, r1.SupFullName } into g2
select new
{
    g2.Key.SupFullName,
    Volunteers = from v0 in g2
                 where v0.VolId != g2.Key.SupId
                 select new { v0.VolFullName, v0.VolPeopleCount }
}

也就是说，我只希望在 Volunteers 中那些 Id 与组 key 不同的人。我期待驱动程序生成一个条件 $push，但它给出了一个简单的:

"$push" : { 
    "VolFullName" : "$VolFullName", 
    "VolPeopleCount" : "$VolPeopleCount" 
}

现在，通过阅读这篇 post我能够将上述查询转换为以下查询:

"$push": {
    "$cond": [
        { "$ne": [ "$VolId", "$SupId" ] },
        { 
            "VolFullName": "$VolFullName",
            "VolPeopleCount": "$VolPeopleCount"
        },
        null
     ]
 }

这给了我想要的结果。我想知道如何(如果可能的话)指示 csharp 驱动程序生成如上所述的查询。提前致谢!

编辑:按照@mickl 的建议，我能够生成一个有条件的$push(还有一个有条件的$sum)但是结果不如预期。假设我的收藏中有以下文档:

{ N: "P_1", S: { I: "S_1", FN: "S_1" }, C: { I: "S_1", FN: "S_1" } }
{ N: "P_2", S: { I: "S_1", FN: "S_1" }, C: { I: "S_1", FN: "S_1" } }
{ N: "P_3", S: { I: "S_1", FN: "S_1" }, C: { I: "S_1", FN: "S_1" } }
{ N: "P_4", S: { I: "S_1", FN: "S_1" }, C: { I: "C_2", FN: "C_2" } }
{ N: "P_5", S: { I: "S_1", FN: "S_1" }, C: { I: "C_2", FN: "C_2" } }
{ N: "P_6", S: { I: "S_1", FN: "S_1" }, C: { I: "C_2", FN: "C_2" } }
{ N: "P_7", S: { I: "S_1", FN: "S_1" }, C: { I: "C_3", FN: "C_3" } }
{ N: "P_8", S: { I: "S_1", FN: "S_1" }, C: { I: "C_3", FN: "C_3" } }
{ N: "P_9", S: { I: "S_1", FN: "S_1" }, C: { I: "C_3", FN: "C_3" } }

这里的N, I, FN, S and C 表示名字, id, 全名, supervisor 和 creator 分别。如您所见，主管 S_1 正在监督每个文档，创建了 3 个文档，另外 2 个创建者创建了 6 个文档，它们也由 S_1 监督。我想做的是由主管总结这些文档，然后由创建者总结文档(人)的数量。所以，我的 LINQ 查询如下所示:

var q = from p in col.AsQueryable()
        group p by new
        {
           SupFullName = p.S.FN,
           SupId = p.S.I,
           CtorFullName = p.C.FN,
           CtorId = p.C.I
        } into g1
        select new
        {
            g1.Key.SupFullName,
            g1.Key.SupId,
            VolFullName = g1.Key.CtorFullName,
            VolId = g1.Key.CtorId,
            VolPeopleCount = g1.LongCount()
        } into r1
        group r1 by new { r1.SupId, r1.SupFullName } into g2
        select new
        {
           g2.Key.SupFullName,
           g2.Key.SupId,
           Volunteers = g2.Select(v => v.VolId != g2.Key.SupId
                           ? new { v.VolId, v.VolFullName, v.VolPeopleCount }
                           : null),
          SupPeopleCount = g2.Select(v => v.VolId == g2.Key.SupId 
                               ? v.VolPeopleCount 
                               : 0).Sum()
        } into r2
        orderby r2.SupPeopleCount descending
        select r2;

查询给我的输出是:

Supervisor: S_1, PeopleCount: 0
   Creator: C_3, PeopleCount: 3
   Creator: C_2, PeopleCount: 3
   Creator: S_1, PeopleCount: 3

这不是我想要的。正确的输出是

Supervisor: S_1, PeopleCount: 3
   Creator: C_3, PeopleCount: 3
   Creator: C_2, PeopleCount: 3

它以下列方式生成条件聚合器(如果我删除条件中的前缀 $_id 并只留下 $SupId 它就像一个魅力):

"__agg0": {
      "$push": {
         "$cond": [
             { "$ne": [ "$VolId", "$_id.SupId" ] },
             {
                "VolId": "$VolId",
                "VolFullName": "$VolFullName",
                "VolPeopleCount": "$VolPeopleCount"
             },
             null
         ]
      }
},
"__agg1": {
    "$sum": {
       "$cond": [
           { "$eq": [ "$VolId", "$_id.SupId" ] },
           "$VolPeopleCount",
           NumberLong(0)
       ]
   }
}

谢谢!

Answer 1

您可以运行 Select 并在 LINQ 语句中使用三元运算符:

group r1 by new { r1.SupId, r1.SupFullName } into g2
select new
{
    g2.Key.SupFullName,
    Volunteers = g2.Select(x => x.VolId != x.SupId ? new { x.VolFullName, x.VolPeopleCount }: null)  
};

这被转化为以下聚合管道:

{ 
    "$group" : { 
        "_id" : { "SupId" : "$SupId", "SupFullName" : "$SupFullName" }, 
        "__agg0" : { 
            "$push" : { 
                "$cond" : [
                        { "$ne" : ["$VolId", "$SupId"] }, 
                        { "VolFullName" : "$VolFullName", "VolPeopleCount" : "$VolPeopleCount" }, 
                        null
                    ] 
                } 
            } 
        } 
}, 
{ "$project" : { "SupFullName" : "$_id.SupFullName", "Volunteers" : "$__agg0", "_id" : 0 } }