我尝试使用C#聚合框架将下一个SQL查询重写为mongodb,但我不明白该怎么做。我需要结合一些结果。
SELECT TOP 100 RES.Agent, RES.Type, RES.Opens FROM ((SELECT UA.ClientDomain AS Agent, UA.Type AS Type, COUNT(*) AS Opens
FROM tReadConfirm AS RC INNER JOIN tUserAgent AS UA ON RC.UserAgentId = UA.Id
WHERE RC.UserId = 2654 AND RC.CampaignId = 27442 AND UA.Type = 1
GROUP BY UA.ClientDomain, UA.Type)
UNION
(SELECT UA.Family AS Agent, UA.Type AS Type, COUNT(*) AS Opens
FROM tReadConfirm AS RC INNER JOIN tUserAgent AS UA ON RC.UserAgentId = UA.Id
WHERE RC.UserId = 2654 AND RC.CampaignId = 27442 AND UA.Type <> 1
GROUP BY UA.Family, UA.Type)) AS RES
ORDER BY RES.Opens DESC
这是我的开始代码,但它不能满足我的需求
db.analytics.aggregate(
{
$match: { UserId: 4749, CampaignId: 93178}
},
{
$group :
{
_id :
{
"Family" : "$userAgent.Family",
"Type" : "$userAgent.Type",
"ClientDomain" : "$userAgent.ClientDomain",
} ,
"Opens":
{
$sum : 1
}
}
},
{$sort :{"Opens":-1}}
)
答案 0 :(得分:4)
我找到了答案,这是我的工作代码示例。运营商&#34; $ cond&#34;帮助我。
db.analytics.aggregate(
{
$match : {
"UserId" : 4790,
"CampaignId" : 93178}
},
{
$group :
{
_id :
{
"Type" : "$userAgent.Type",
"ClientDomain" : { $cond: { if: { $eq: [ "$userAgent.Type", 1 ] }, then: "$userAgent.ClientDomain", else: "$userAgent.Family" }},
} ,
"Opens":
{
$sum : 1
}
}
},
{
$sort :{"Opens":-1}
},
{
$limit:10
}
)
答案 1 :(得分:0)
您可以尝试以下想法:
通过$ project运算符获取新的文档结构。像
这样的东西{剂: [{Agent:“$ userAgent.Family”,Index:0},{Agent:“$ userAgent.ClientDomain”,Index:1}],类型:“$ userAgent.Type”}
$展开代理字段,这样您就可以获得2个文档而不是1个包含2个元素数组的文档。
使用以下想法重新过滤:{$或:[{Index:0,Type:1},{Index:1,Type&lt;&gt; 1}]}
按代理,类型和索引字段分组。
如果我没弄错的话,你会得到同样的结果。
编辑:不幸的是,第2步无效。 MongoDB: how create a new array field with the aggregate framework