Log file Viewer
#mongodb logs for Saturday the 19th of April, 2014
[01:37:24] <Ephexeve> Hey guys, is there a way to put a password to a database? So when I would like to query anything from that db, I need a password first? Otherwise there is NO way I can access it?
[03:19:22] <Reptar> How do I create a collection that is comprised of aggregate groupings from other collections?
[03:40:21] <Mark_> db.commandlog.distinct('Account', { AccessLevel: { $ne: 'Player', $ne: 'Administrator', $ne: 'Developer', $ne: 'Owner' } } );
[03:57:20] <laprice> but a. that seems really lame. b. http://docs.mongodb.org/manual/reference/method/cursor.explain/#explain.millis
[03:58:37] <laprice> So I'm bashing away at system.profile.find() and I want to always get the most recent query. But I am not getting any joy.
[03:59:56] <laprice> system.profile.find() doesn't let you do system.profile.find({ ns : "collname" }) and only get info on queries against your collection?
[04:07:09] <laprice> system.profile.fine({ ns : "dbname.collname" }).limit(1) goes in the wrong order
[04:15:32] <laprice> http://docs.mongodb.org/manual/reference/method/cursor.explain/#explain.millis is that just a doc bug? b/c "reflects" does not tell me if that number is milliseconds, or if it's some metric vaguely related to, or derived from the number of milliseconds required to find the document.
[09:50:47] <tinco> hey guys, I'm performing an update, and in this update I want to upsert an element but in the documentation there is only a push, and the push doesn't seem to have an upsert modifier
[09:54:23] <tinco> so I have this collection: {"notifications" : [{"id" : 1, "occurrences" : [{"id" : 1, "created_at" : 353251}]}]}
[09:55:26] <tinco> so I have this collection: {"notifications" : [{"id" : 1, "occurrences" : [{"id" : 1, "created_at" : 353251}], "uniqueProperty" => "superUnique"}]}
[09:58:22] <tinco> and then I do notifications.update({"uniqueProperty" : "superUnique"}, {'$push' : {'occurrences' : {"created_at" : 353251 }}, {"upsert" => true})
[09:59:14] <tinco> but this will push a new occurrence every time it gets updated, I only want to push a new occurrence if the created_at does not match an existing occurrence
[10:04:38] <tinco> so I create a new notification, or update an existing one based on "uniqueProperty", and when I update an existing one, I want to add to its occurrences array, but only if there's no occurrence at that date yet for that notification
[10:07:03] <kali> notifications.update({"uniqueProperty" : "superUnique", "occurencies.created_at" : {$ne: 353251} }, {'$push' : {'occurrences' : {"created_at" : 353251 }}, {"upsert" => true})
[10:09:16] <tinco> won't that create a new notification whenever there's a new occurrence, even if it has the same uniqueProperty?
[10:10:32] <tinco> I realize my query is a bit absurd, because I'm basically doing 2 upserts in one query
[10:12:11] <tinco> so this query is in a pretty time critical location, it might be called hundreds of times per second, so I'd like to avoid making it two queries
[10:14:16] <tinco> my concern is that there might be hundreds of occurrences per second, and if I just push them all into that occurrences array, it could become very big and slow to query, or perhaps even go over the mongo element size limit
[10:28:43] <tinco> kali: http://privatepaste.com/c2799446b5 <-- see how it creates a second notification object, when the created at matches? or did I make a mistake?
[10:30:56] <tinco> yeah that's the problem, I can't use a unique index, I don't know the unique index yet when I create the notification
[10:35:54] <Zelest> kali, when I have this array later on, {_id: ObjectId, to: [links1, links2, links3, ...]} .. can I efficiently get the number of links for let's say links3
[10:44:50] <tinco> kali: is it possible to get the update result printed in the mongo shell? I can't find an option anywhere
[15:53:47] <ThiefMaster> hi, i'm trying to search documents using this condition: "somefield needs to contain all words from the search query". is there a nice way to do this using mapreduce or the aggregation pipeline? or should i use fulltext search instead?
[15:54:19] <ThiefMaster> (meh, fulltext-search is not an option since gentoo doesn't have an ebuild package for 2.6 :x)
[20:03:25] <Jadenn> depending on what language you are accessing it from, you would get the root array
[21:01:03] <Mark_> { "_id" : ObjectId("5352df61fbaa88b24504a810"), "480" : { "999" : { "0" : "587F", "-1" : "579F" } } }