Consider the following dataset:
(考虑以下数据集:)
db.dum.insert({ownerId:1, requestId:'a', price:3, createdAt:3, isWinner:true})
db.dum.insert({ownerId:1, requestId:'a', price:1, createdAt:1, isWinner:false})
db.dum.insert({ownerId:1, requestId:'a', price:2, createdAt:2, isWinner:true})
db.dum.insert({ownerId:1, requestId:'b', price:4, createdAt:2, isWinner:true})
db.dum.insert({ownerId:1, requestId:'b', price:5, createdAt:1, isWinner:false})
db.dum.insert({ownerId:2, requestId:'b', price:5, createdAt:1, isWinner:false})
You could use $reduce
(您可以使用$ reduce)
Here, for a grouping id, we keep all matching documents as an array ( candidates
).
(在这里,对于分组ID,我们将所有匹配的文档保留为数组( candidates
)。)
On the project
stage, for each group we iterate through the array, and reduce it to the minimal element found (by price that is)
(在project
阶段,对于每个组,我们迭代数组,并将其减少到找到的最小元素(按价格计算))
db.dum.aggregate([{
$match: {
ownerId: 1
}
},
{
$group: {
_id: "$requestId",
candidates: { $push:'$$ROOT'}
}
},
{
$project:{
item: {
$reduce: {
input: '$candidates',
initialValue: '$candidates.0',
in: {
$cond: {
if: {
$lt: ['$$value.price', '$$this.price']
},
then:'$$value',
else:'$$this'
}
}
}
}
}
},
{
$replaceRoot:{newRoot:'$item'}
}
]).toArray()
output:
(输出:)
[
{
"_id" : ObjectId("5ddcc8e0eb1f0217802fb507"),
"ownerId" : 1,
"requestId" : "b",
"price" : 4,
"createdAt" : 2,
"isWinner" : true
},
{
"_id" : ObjectId("5ddcc8e0eb1f0217802fb505"),
"ownerId" : 1,
"requestId" : "a",
"price" : 1,
"createdAt" : 1,
"isWinner" : false
}
]