Select latest documents in MongoDB - mongodb

Question

My app records some data in MongoDB collection. There are three types of events: 'event1', 'event2', 'event3'. And this elements adds periodically.
Event structure example:
{ 'Data' : 'a234235', 'Type' : 'event1', 'Timestamp' : 1366006599 }
{ 'Data' : 'b978543', 'Type' : 'event2', 'Timestamp' : 1366006600 }
{ 'Data' : 'c567921', 'Type' : 'event3', 'Timestamp' : 1366006601 }
{ 'Data' : 'd327863', 'Type' : 'event1', 'Timestamp' : 1366007100 }
{ 'Data' : 'e012315', 'Type' : 'event2', 'Timestamp' : 1366007102 }
{ 'Data' : 'f834721', 'Type' : 'event3', 'Timestamp' : 1366007103 }
Please help compose right query for get current state of each event in database. I need three elements of different types and maximal timestamp.

Answer 1

Since you may have ~200 events (as noted in your comments on the question), I would suggest the most efficient approach would be to create a summary document that records the latest events (by type). This avoids the needs to do separate queries and should be easy to maintain using a $set as new event types are observed.
The summary doc would look like:
{
'Event1' : { 'Data' : 'd327863', 'Timestamp' : 1366007100 },
'Event2' : { 'Data' : 'e012315', 'Timestamp' : 1366007102 },
'Event3' : { 'Data' : 'f834721', 'Timestamp' : 1366007103 }
}
This pre-aggregated report pattern will avoid the need for multiple queries if you often need to find the maximal event by type, and save on (potentially large) space for an index on Timestamp which would otherwise be needed to find the latest entry efficiently.

Answer 2

You can use the following JavaScript to get your work done. The drawback here is that I am assuming your events are constant else you need to change this in the "Type" array field.
You can invoke this while connecting to DB.
a=function()
{
var Type = new Array();
Type[0]="event1";
Type[1]="event2";
Type[2]="event3";
Type[3]="event4";
Type[4]="event5";
for (i=0;i<Type.length;i++)
{
var myCursor = db.foo.find({Type:Type[i]}).sort({Timestamp:-1}).limit(1);
myCursor.forEach(printjson);
}
}
and this is the output I got for my test case.
MongoDB shell version: 2.2.2
connecting to: test
{
"_id" : ObjectId("516bc991cde2925693705103"),
"Data" : "d327863",
"Type" : "event1",
"Timestamp" : 1366007100
}
{
"_id" : ObjectId("516bc991cde2925693705104"),
"Data" : "e012315",
"Type" : "event2",
"Timestamp" : 1366007102
}
{
"_id" : ObjectId("516bc992cde2925693705105"),
"Data" : "f834721",
"Type" : "event3",
"Timestamp" : 1366007103
}
bye
Also note that, by default it connects to test db. Indexing on Type and Timestamp will help you in performance.

Answer 3

You will need a separate query for each event.
db.yourCollection.find({Type:"event1"}).sort({Timestamp:-1}).limit(1);
db.yourCollection.find({Type:"event2"}).sort({Timestamp:-1}).limit(1);
db.yourCollection.find({Type:"event3"}).sort({Timestamp:-1}).limit(1);
Note that creating an index on the Timestamp field will likely improve the performance a lot.
See also: http://docs.MongoDB.org/manual/reference/method/cursor.sort/