Convert query using Spring Mongodb Aggregation apis - spring

Question

db.flm_conversation.aggregate(
[
{$match: {"conversationRecords.isPrimary":true,"conversationRecords.commentTime":{'$gte': new Date('2016-12-18 00:00:00'), '$lte': new Date('2016-12-18 23:59:59')} } },
{$unwind:"$conversationRecords"},
{ $group: {
_id: {
"commentLevel": "$commentLevel",
"time":{"$add": [
{
"$subtract": [
{ "$subtract": [
"$conversationRecords.commentTime",
new Date(0)
]}
,
{ "$mod": [
{ "$subtract": [
"$conversationRecords.commentTime",
new Date(0)
]},
1000 * 60 * 30
]}
]
},
new Date(0)
]}
},
count: { "$sum": 1 }
}},
{ $group: {_id: "$_id.commentLevel",count: { "$sum": 1 },pointrecord:{$push: {time:"$_id.time",count:"$count"} } }},
{ $project: { _id: 1,count:1,pointrecord:1 } }
])
How to convert this query using Spring Mongodb Aggregation apis?
AggregationOperation match = Aggregation.match(Criteria.where("conversationRecords.isPrimary").is(true)
.and("conversationRecords.commentTime").gte(DateUtils.stringToDate("2016-12-18 00:00:00","yyyy-MM-dd HH:mm:ss")).lte(DateUtils.stringToDate("2016-12-18 23:59:59","yyyy-MM-dd HH:mm:ss")));
AggregationOperation unwind = Aggregation.unwind("conversationRecords");
AggregationOperation group = Aggregation.group("commentLevel");
Aggregation agg = newAggregation(match,unwind,group);
AggregationResults<SummaryRecordAggre> groupResults
= mongoTemplate.aggregate(agg, COLLECTION_NAME, SummaryRecordAggre.class);
I don't know group "$add" how to convert?I find by
http://docs.spring.io/spring-data/mongodb/docs/current/reference/html/#mongo.aggregation

Answer 1

You can try something like this. I removed the extra operators which I felt was not needed but we can easily put them back should you need.
The first group which is something you need to build using MongoDB objects as currently all aggregation operators are not supported out of box in spring mongo.
Similar way for push operator in second group as you cant push multiple values.
AggregationOperation match = Aggregation.match(Criteria.where("conversationRecords.isPrimary").is(true)
.and("conversationRecords.commentTime").gte(DateUtils.stringToDate("2016-12-18 00:00:00", "yyyy-MM-dd HH:mm:ss")).lte(DateUtils.stringToDate("2016-12-18 23:59:59", "yyyy-MM-dd HH:mm:ss")));
AggregationOperation unwind = Aggregation.unwind("conversationRecords");
AggregationOperation firstGroup = context -> context.getMappedObject(new BasicDBObject(
"$group", new BasicDBObject(
"_id", new BasicDBObject("commentLevel", "$commentLevel")
.append(
"time", new BasicDBObject(
"$subtract", new Object[]{"$conversationRecords.commentTime", new BasicDBObject(
"$mod", new Object[]{"$conversationRecords.commentTime", 1000 * 60 * 30})}))).append("count", new BasicDBObject("$sum", 1))));
AggregationOperation secondGroup = Aggregation.group("_id.commentLevel").count().as("count").push(new BasicDBObject("time", "$_id.time").append("count", "$count")).as("pointrecord");
AggregationOperation project = Aggregation.project("_id", "count", "pointrecord");
Aggregation agg = newAggregation(match, unwind, firstGroup, secondGroup, project);
This is equivalent of below query
[{
"$match": {
"conversationRecords.isPrimary": true,
"conversationRecords.commentTime": {
"$gte": {
"$date": "2016-12-27T14:46:50.896Z"
},
"$lte": {
"$date": "2016-12-27T14:46:50.896Z"
}
}
}
}, {
"$unwind": "$conversationRecords"
}, {
"$group": {
"_id": {
"commentLevel": "$commentLevel",
"time": {
"$subtract": ["$conversationRecords.commentTime", {
"$mod": ["$conversationRecords.commentTime", 1800000]
}]
}
},
"count": {
"$sum": 1
}
}
}, {
"$group": {
"_id": "$_id.commentLevel",
"count": {
"$sum": 1
},
"pointrecord": {
"$push": {
"time": "$_id.time",
"count": "$count"
}
}
}
}, {
"$project": {
"_id": 1,
"count": 1,
"pointrecord": 1
}
}]
}