Log file Viewer
#mongodb logs for Monday the 19th of January, 2015
[00:12:16] <jumpman> Is there a way to structure an .update so I can use $addToSet on a field an object may or may not have?
[00:12:35] <jumpman> Currently: DAL.UsersDAO.collection.update({_id: userId}, {$addToSet: {roles: permission}});
[00:24:22] <jayjo> so I'll use the one-to-many relationship with document references, but if I go this way I'll be keeping objects connected through the objectid, should I create a separate id variable to be used in my url and keep the reference through objectid?
[02:29:42] <huleo> this surely will be simple - I'm taking my first steps with aggregation framework
[02:30:41] <huleo> which is cool - it takes "query" argument, to actually use it when querying for documents to process
[02:31:19] <huleo> but sometimes I'm not querying by geo-location...how will I "find" documents by basic, simple query, to pass further to the pipeline?
[02:43:20] <huleo> [{"$match":{"something":{"$elemMatch":{"smthelse":{"$gt":"2005-07-17T20:23:22.247Z"}}}}}]
[02:44:23] <joannac_> huleo: can you pastebin a document that matches with a find(), but not the aggregation $match?
[02:55:56] <joannac_> huleo: yeah, your elemMatch works for me here. I was about to ask why you're storing as strings rather than Date objects
[02:56:49] <huleo> joannac_: it's something like "publishedSince" filter, value of it given as number of seconds
[03:03:34] <sgo11> hi, I just want to log all user queries. I did "db.setProfilingLevel(2)". When I do "db.system.profile.find().pretty()", it shows there are many queries against "test.system.indexes" and "test.system.profile" (I don't know what triggers those queries). after a few minutes, 400 queries against "indexes" and "profile" ns compared to only 7 queries against my actual collection. how can I disable logging "indexes" and "profile" queries?
[03:03:42] <huleo> $project - will I actually be able to /exclude/ specific fields, instead of including them one-by-one?
[03:16:04] <TheAncientGoat> Being lazy here, but does anyone have an estimate of the perf/efficiency benefits of using the aggregation pipeline over just manually grouping fetched data in node for a datataset of a couple k docs?
[03:28:21] <huleo> TheAncientGoat: all I can speak about is my case - and here aggregation saves me plenty of tinkering to get distance of each result (2dsphere) from the point I'm querying against
[03:28:56] <Boomtime> huleo: that sounds like you are benefitting from the geo API not strictly the aggregation part
[03:31:20] <TheAncientGoat> My case is basically grouping transactions by day, a lot less LOC through aggregation, but not sure what the perf considerations are regarding it
[03:33:09] <Jonno_FTW> is it wise to have very large documents? for example, my data looks like: site_number: {date:[{2010-01-01:[{time:10:00,readings:[{detector:1,count:50}]}]}]}}, where there is ~2m individual readings, spread across different sites,dates and times
[03:35:04] <Boomtime> huleo: I assume you mean you specifically use "distanceField" in your pipeline, although that is aggregation specific, it is the geo engine that supplies the value and does the calculation - for you, this does mean that aggregation has significant convenience over doing the calcs client-side
[03:36:03] <Boomtime> clearly however, the calculation of distance is performed in regular query since $geoNear doies precisely this - you just don't have access to the value
[03:40:21] <Boomtime> TheAncientGoat: there are still lots of variables, and it depends also on what you actually desire to "optimize" - do you want to save network bandwidth or server resources?
[03:44:27] <TheAncientGoat> I have a lot of concurrent, low bandwidth queries for other data. Not sure if I'll be choking at bandwidth or CPU first, but the connections to the clients could be pretty slow, so both are a factor. Aggregation will save bandwidth, I can see that, I just don't know at what cost
[03:46:09] <Boomtime> you're not going to get a definitive answer, it depends on CPU, memory, disk IO (current load of previous three, and how much spare), how many documents are involved, how frequently this is run...
[03:47:57] <Boomtime> if you match, sort and group on fields which are indexed, then the pipeline can take advantage of that - something the remote client cannot - but the difference might not matter, depending on the numerous variables I mentioned before
[03:49:36] <Boomtime> btw, it's not just bandwidth, it's also latency - if you have more than a few hundred documents, the cursor has to go back to the server to get more - if you group these, you send less and subsequently get fewer round-trips to the server
[04:15:15] <Nilium> I'm just maintaining the awful code that produced said documents, and it makes me want to do a sick flip off a building.
[09:46:08] <Derick> db.colname.update( { _id: ... }, { '$set' : { 'fieldname.subfieldname' : 42 } } ); ought to do... it's here in the docs: http://docs.mongodb.org/manual/reference/method/db.collection.update/#update-parameter and http://docs.mongodb.org/manual/core/document/#document-dot-notation
[10:40:04] <nelasx> hi! my mongodb DB got big, is there a way to shrink the white space and reclaim HDD space?
[10:44:59] <nelasx> kali: i will just extend the lvm, and latter in the week dump the data to a new replica set
[10:46:13] <nelasx> kali: will these work out? http://blog.mongolab.com/2014/01/managing-disk-space-in-mongodb/
[10:52:18] <sjose_> kali, I want to update an embedded document using its Id is it possible using $set...?
[11:00:17] <kali> sjose_: db.blah.update({ "trainings.id" : ...}, { $set: { "trainings.$.date" : ... }})
[11:00:59] <StephenLynx> for you to use the $ on the update block, you would have to also have something to query on the math block
[11:02:21] <sjose_> kali, StephenLynx : as trainings is a embedded document how can I locate that first...?
[11:20:04] <sjose_> kali, If I want to update the embedded item as a whole using its id ..how shall I do it...???
[12:35:16] <mnms_> Im trying to sum field for all my documents in collection. It takes much time. So I added single field index on this field, but it also didnt help
[12:36:11] <kali> yeah, nothing will make it significantly better. your best option is to denormalize
[12:37:00] <mnms_> kali: what does it means denormalize ? Cause for me it cannot be more denormalized..
[12:47:37] <mnms_> and my question is is there any way to check if query will use index without waiting til its over
[14:37:21] <sjose_> StephenLynx, Please have a look over here 'http://justpaste.it/iz7v', here 'trainings' is an embedded document. so can I update trainings with "_id" : ObjectId("54bcaf4eba09015284646f6d") with all new values in an easier way....?
[16:25:54] <Raffaele> Hello. I've noticed there seems to be some mismatch between engine_v8-3.25.cpp vs engine_v8.cpp up to the point that the former doesn't even compile because of silly errors like string and set missing std::
[17:07:54] <toter> Hi everybody... I'm running a collection against another to find missing ['company'] values. The first collection has 345 documents and the second one, 20285. This operation is taking 63 seconds to complete... Is there anything I can do to improve this execution time? Code: http://hastebin.com/ifayexesuv
[19:49:16] <sekyms> I have a question regarding Schema as it relates to this: http://docs.mongodb.org/ecosystem/use-cases/product-catalog/#schema
[22:35:14] <FunnyLookinHat> Is it fairly trivial to move a 2/1 ( member/arbiter ) set to a 3/0 ( primary with two secondary members ) ?
[23:10:13] <mortal1> howdy folks, say have a collection foo. this collection has a document with an id of 1, and a object of bar that I'd like to remove
[23:34:25] <liamkeily> I want to append to an array in mongo. But want to make sure its not overwritten by another process. Whats the recommended way to do this?