I am using mongodb as a document oriented database, and spring data as the ODM with it. I am facing hard time, performing a max aggregation on complex bson structure.
I have to find the max date, from all documents but if the document has an embedded document, it has to consider that embedded document for the max date.
Here is an example, lets suppose i have a collection name person and person collection contains following documents.
{
"_id" : ObjectId("55def1ceb5b5ed74ddf2b5ce"),
"name" : "abc",
"birth_date_time" : '15 June 1988'
"children" : {
"_id" : ObjectId("55def1ceb2223ed74ddf2b5ce"),
"name" : "def",
"birth_date_time" : '10 April 2010'
}
},
{
"_id" : ObjectId("55def1ceb5b5ed74dd232323"),
"name" : "xyz",
"birth_date_time" : '15 June 1986'
},
{
"_id" : ObjectId("55def1ceb5b5ed74ddf2b5ce"),
"name" : "mno",
"birth_date_time" : '18 March 1982'
"children" : {
"_id" : ObjectId("534ef1ceb2223ed74ddf2b5ce"),
"name" : "pqr",
"birth_date_time" : '10 April 2009'
}
}
It should return 10 April 2010 as this the max birth date for a person in the collection person. I want to know who to achieve it using spring data repository.
Here are the MongoDB aggregations. They should be easily implemented in Spring Data.
db.person.aggregate([
{$group: {
_id: null,
maxDate: {$max : {
$cond: [
{$gt : ["$birth_date_time","$children.birth_date_time"]},
"$birth_date_time",
"$children.birth_date_time"
]}}
}}
])
or using a $project:
db.person.aggregate([{
$project: {
mDate: {
$cond: [
{$gt : ["$birth_date_time","$children.birth_date_time"]},
"$birth_date_time",
"$children.birth_date_time"
]
}
}},
{$group: {
_id: null,
maxDate: {$max : "$mDate"}
}},
])
Related
In Java with MongoDB driver I want to have pairs (business_name, count), meaning reviews count per business. Currently I have aggregate pipeline:
Bson group = group("$business_id", Accumulators.sum("count", 1));
Bson lookupOperation = lookup(
"business",
"_id",
"business_id",
"business_data"
);
Bson unwind = unwind("$business_data");
Bson project = project(fields(include("business_data.name", "count"), excludeId()));
return db
.getCollection("tip")
.aggregate(Arrays.asList(group, lookupOperation, unwind, project));
It works, but returns:
Document{{count=1, business_data=Document{{name=Firestone Complete Auto Care}}}}
1) How can I unwind business_data.name to have {count, name}?
2) How can I make name distinct? I want only 1 count per name, but printing results gives many identical copies, e. g.:
Document{{count=3, business_data=Document{{name=Malmaison}}}}
Document{{count=3, business_data=Document{{name=Malmaison}}}}
Results are quite big collection, so I return AggregateIterable, but I want results sorted by name. Can I do that with iterable, without loading entire data into array and sorting the array?
EDIT
Business document example:
{
"_id" : ObjectId("5ddbc3c1a94f7aac8d179b7c"),
"business_id" : "vcNAWiLM4dR7D2nwwJ7nCA",
"full_address" : "4840 E Indian School Rd\nSte 101\nPhoenix, AZ 85018",
"hours" : {
"Tuesday" : {
"close" : "17:00",
"open" : "08:00"
},
"Friday" : {
"close" : "17:00",
"open" : "08:00"
},
"Monday" : {
"close" : "17:00",
"open" : "08:00"
},
"Wednesday" : {
"close" : "17:00",
"open" : "08:00"
},
"Thursday" : {
"close" : "17:00",
"open" : "08:00"
}
},
"open" : true,
"categories" : [
"Doctors",
"Health & Medical"
],
"city" : "Phoenix",
"review_count" : 7,
"name" : "Eric Goldberg, MD",
"neighborhoods" : [],
"longitude" : -111.983758,
"state" : "AZ",
"stars" : 3.5,
"latitude" : 33.499313,
"attributes" : {
"By Appointment Only" : true
},
"type" : "business"
}
Review document example:
{
"user_id": "IORZRljfUkedhh1SGMthTA",
"text": "The desserts are enormous..dear god.",
"business_id": "JwUE5GmEO-sH1FuwJgKBlQ",
"likes": 0,
"date": "2011-09-29",
"type": "tip"
}
I'm looking for a way to implement graphlookup using Java in a Resful Web API. I'm trying to implement the hierarchy like on the MongoDB (https://docs.mongodb.com/v3.4/reference/operator/aggregation/graphLookup/)
{ "_id" : 1, "name" : "Dev" }
{ "_id" : 2, "name" : "Eliot", "reportsTo" : "Dev" }
{ "_id" : 3, "name" : "Ron", "reportsTo" : "Eliot" }
{ "_id" : 4, "name" : "Andrew", "reportsTo" : "Eliot" }
{ "_id" : 5, "name" : "Asya", "reportsTo" : "Ron" }
{ "_id" : 6, "name" : "Dan", "reportsTo" : "Andrew" }
This is the employee collection, what I want is to be able to create this structure stored in MongoDB
{
"_id" : 1,
"name" : "Dev",
"reportingHierarchy" : [ ]
}
{
"_id" : 2,
"name" : "Eliot",
"reportsTo" : "Dev",
"reportingHierarchy" : [
{ "_id" : 1, "name" : "Dev" }
]
}
{
"_id" : 3,
"name" : "Ron",
"reportsTo" : "Eliot",
"reportingHierarchy" : [
{ "_id" : 1, "name" : "Dev" },
{ "_id" : 2, "name" : "Eliot", "reportsTo" : "Dev" }
]
}
I've seen examples like this for aggregation, but nothing on graphlookup
Aggregation agg = newAggregation(
match(Criteria.where("pageId").is("2210")),
unwind("postIds"),
group("_id").sum("1").as("sum")
//project("$sum").and("pageId").previousOperation()
);
Is there a way to get graphlookup into a format like this? Where instead of using match, unwind, group, I can use GraphLookupOperation and then something like get map result.
AggregationOperation aggregationOperation = new AggregationOperation() {
#Override public DBObject toDBObject(AggregationOperationContext aggregationOperationContext) {
DBObject graphLookup = new BasicDBObject(
"from", "individual").append(
"startWith", "$reportsTo").append(
"connectFromField", "reportsTo").append(
"connectToField", "firstName").append(
"maxDepth", 2).append(
"as", "reportingHierarchy");
return new BasicDBObject("$graphLookup", graphLookup);
This code allows for a workaround since i'm using an older version, I think 1.10.0 spring mongo. Now I have the issues of the "reportingHierarchy" not looking like I want it to. It's not giving me a name. Just reportsTo and reportingHierarchy that is also including _class which I don't want.
Hi I am using spring data mongo, I need to fetch data based on multiple where condition. The problem I have when I want to apply a where clause to a list in a list.
For example
{
"_id" : ObjectId("5982bf9339f3c92b84be4737"),
"_class" : "com.paladion.payment.model.GroupQuestionMapping",
"saqID" : "SAQ A",
"saqVersion" : "3",
"questionTab" : {
"Secure Network" : [
{
"number" : "2.1 (a)",
"question" : "Are vendor-supplied"
"description" : "<ul><li>Review"
},
{
"number" : "2.1 (b)",
"question" : "Are unnecessary"
"description" : "<ul><li>Review policies
}
],
"Access Control" : [
{
"number" : "2.1 (a)",
"question" : "Are vendor-supplied"
"description" : "<ul><li>Review"
},
{
"number" : "2.1 (b)",
"question" : "Are unnecessary"
"description" : "<ul><li>Review policies
}
]
}
}
Over here I need to fetch data where saqId is SAQ A and saq Version is 3 and questionTab is secure network.
I have problem in applying criteria on questionTab.
my code:
Query query = new Query();
query.addCriteria(Criteria.where("saqtype").is(saqType));
query.addCriteria(Criteria.where("saqversion").is(saqVersion));
query.addCriteria(/* criteria on questionTab */);
query.addCriteria(Criteria.where("questionTab.Secure Network").exists(true));
Thing to note is that it will bring you the full document based on criteria, so you would have to filter out other type of questionTab from document.
Other way is aggregation but then I think processing on application layer might be preferable.
I have a collection containing following documents -
{
"_id" : ObjectId("56e7a51b4a66e30330151847"),
"host" : "myTestHost.com",
"sessionId" : "daxxxxxx-xxxx-xxxx-xxxx-xxxxxxx1",
"ssoId" : "xxxxxx#gmail.com",
"days" : {
"13" : NumberLong(130),
"11" : NumberLong(457),
"10" : NumberLong(77)
},
"count" : NumberLong(664),
"timeStamp" : NumberLong("1458021713370")
}
I am using mongo java driver 3.2.1.
This document contains an embedded document 'days', that holds a specific count for each day of month.
I need to find the number of days for which a count is present.For example - for above document mentioned, the number of days for which count is present is 3 (13th, 11th and 10th day of month).
I know how to get the count on mongo console -
mongos>var count = 0;
mongos> db.monthData.find({},{days:1}).forEach(function(record){for(f in record.days) { count++;}});
mongos> count;
I need to convert this to java code.
Maybe you can reshape your schema as follow:
{
"_id" : ObjectId("56e7a51b4a66e30330151847"),
"host" : "myTestHost.com",
"sessionId" : "daxxxxxx-xxxx-xxxx-xxxx-xxxxxxx1",
"ssoId" : "xxxxxx#gmail.com",
"days" : [
{
"13" : NumberLong(130)
},
{
"11" : NumberLong(457)
},
{
"10" : NumberLong(77)
}
],
"count" : NumberLong(664),
"timeStamp" : NumberLong(1458021713370)
}
days becomes an array of objects and in this way you can easily use aggregation pipeline to know how many elements are in days array:
>db.collection.aggregate(
[
{$match:{days:{"$exists":1}}},
{$project:{
numberOfDays: {$size:"$days"},
_id:1}
}
]
)
The aggregation returns:
{ "_id" : ObjectId("56e7a51b4a66e30330151847"), "numberOfDays" : 3 }
To use Aggregation Pipeline with Java driver see aggregate, AggregateIterable, Block and read Data Aggregation with Java Driver
I am having following document in mongo,
{
"_id" : ObjectId("506e9e54a4e8f51423679428"),
"description" : "ffffffffffffffff",
"menus" : [
{
"_id" : ObjectId("506e9e5aa4e8f51423679429"),
"description" : "ffffffffffffffffffff",
"items" : [
{
"name" : "xcvxc",
"description" : "vxvxcvxc",
"text" : "vxcvxcvx",
"menuKey" : "0",
"onSelect" : "1",
"_id" : ObjectId("506e9f07a4e8f5142367942f")
} ,
{
"name" : "abcd",
"description" : "qqq",
"text" : "qqq",
"menuKey" : "0",
"onSelect" : "3",
"_id" : ObjectId("507e9f07a4e8f5142367942f")
}
]
},
{
"_id" : ObjectId("506e9e5aa4e8f51423679429"),
"description" : "rrrrr",
"items" : [ {
"name" : "xcc",
"description" : "vx",
"text" : "vxc",
"menuKey" : "0",
"onSelect" : "2",
"_id" : ObjectId("506e9f07a4e8f5142367942f")
} ]
}
]
}
Now , i want to update the following document :
{
"name" : "abcd",
"description" : "qqq",
"text" : "qqq",
"menuKey" : "0",
"onSelect" : "3",
"_id" : ObjectId("507e9f07a4e8f5142367942f")
}
I am having main documnet id: "_id" : ObjectId("506e9e54a4e8f51423679428") and menus id
"_id" : ObjectId("506e9e54a4e8f51423679428") as well as items id "_id" : ObjectId("507e9f07a4e8f5142367942f") which is to be updated.
I have tried using the following query:
db.collection.update({ "_id" : { "$oid" : "506e9e54a4e8f51423679428"} , "menus._id" : { "$oid" : "506e9e5aa4e8f51423679429"}},{ "$set" : { "menus.$.items" : { "_id" : { "$oid" : "506e9f07a4e8f5142367942f"}} , "menus.$.items.$.name" : "xcvxc66666", ...}},false,false);
but its not working...
The positional operator does not work on the number of levels you are trying to get it to work on ( https://jira.mongodb.org/browse/SERVER-831?focusedCommentId=22438&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel ) with menus.$.items.$.name and even if it did MongoDB query parser would have no idea what the other $ is from the find of the update.
You will need to pull out the items from the schema, update that seprately and then update the root document.
One good way of judging when queries should be done separately is to think that each menu sounds like a separate entity (or table in a relational database) as such you should probably work on updating those entites (or tables in a relational model) separately to the parent entity (table).
So first you would get out the main root document. Scroll across it's menus in client side and then $set that particular menu to the entire item you build on client side.
Edit
The way I imagine this work client side is (in pseudo code since my Java is a little rusty) by first getting that document in an active record fashion:
doc = db.col.find({ "_id" : { "$oid" : "506e9e54a4e8f51423679428"} ,
"menus._id" : { "$oid" : "506e9e5aa4e8f51423679429"}});
Then you would iterate through the document assigning your values:
foreach(doc.menus as menu_key => menu){
foreach(menu['items'] as key => item){
if(item._id == { "$oid" : "506e9f07a4e8f5142367942f"}){
doc.menus[menu_key][key][name] = "xcvxc66666"
}
}
}
And then simple save the doc after all changes are commited:
db.col.save(doc);
This is of course just one way of doing it and this way uses the activen record paradigm which I personally like. In this idea you would combine the find with everything else you need to modify on the document, building it up client side and then sending it all down as one single query to your DB.