Log file Viewer
#mongodb logs for Friday the 1st of February, 2013
[00:00:39] <diamonds> so if you have 3 keys in your query, then you change one statement from $set to a straight document replacement
[00:04:09] <diamonds> I guess it's just not intuitive to me that keys/values would be pulled out of the query argument and used as data to insert
[03:13:39] <ronar> Hi all! I have issue with mongo on OS X. I download mongo and there's a bin folder in it. But how to install this folder in a system, because it's very unhandy to navigate to ~/Download folder to exec mongod.
[03:19:58] <Mike1382> Hi, I'm looking at Django and MongoDB framework options and can't seem to open the link related to Django MongoDB Engine
[03:26:11] <arnpro> not really Mike1382, but a quick google search gave me this: https://groups.google.com/d/msg/django-non-relational/qEt4xSjuwYA/REj98YTaDvQJ
[03:48:30] <Mike1382> I am bit confused with regards to Mongo Engine and Mongo non-real.. Are they separate things? Can I use Mongo Engine without a forked version of Django?
[03:50:01] <Mike1382> If they are separate things, then what are the benefits of using the forked non-relational version rather than mongo engine + classic version of Django
[06:28:38] <axilaris> hi... can i get some advise from u guys...im new to mongodb... how should i model 2 tables....should i use belongs_to or embedded_to
[09:04:13] <igotux> i have a small sharder cluster having a databae of few gbs .. am plannig to do mongodump on a mongos... :- http://docs.mongodb.org/manual/tutorial/backup-small-sharded-cluster-with-mongodump/
[09:27:51] <axilaris> just playing around mongodb.... im just curious... is it alright to have different tables with relation to it ? because in mongo, whatever table inserts will just go into the same db....
[09:33:12] <axilaris> everything is just inserted into the same db... u would have to query the relationship to aggregate the data
[09:36:16] <axilaris> true also... u can always query to get the info... a bit strange a bit dealing with a flat table
[09:38:01] <axilaris> ok... u dont have a strong opinion on how to go with... i guess what u r saying... if i want performance and scalability then nosql...otherwise... sql
[09:40:03] <NodeX> only YOU can make decisions about YOUR data because YOU know YOUR data better than me
[09:40:33] <NodeX> personaly I hate relations, they're not efficient so I don't use relational DB's at all
[09:46:20] <NodeX> I have the luxury of making the decisions for my apps/clients so I can go that route
[13:37:42] <arnpro> how can I write a map reduce that groups a set of records by user, but only the first n documents of each user?
[13:53:22] <jefersonleao> hey everybody, can anyone give me a little help about read preferences with java driver??
[13:54:41] <jefersonleao> i run into a problem in that i cannot specify it when perform a simple "find" operation
[13:59:12] <jefersonleao> idk but i think you probably can mapreduce your set with a simple js function
[13:59:18] <kali> arnpro: this channel is usually active, but your question has no satisfactory answer
[14:01:21] <kali> you don't how to use the api ? you get an error ? you get no error but the preference is not honoured ?
[14:02:57] <bean|work> Deprecated. As of release 2.9.0, replaced by ReadPreference.secondaryPreferred(DBObject firstTagSet, DBObject... remainingTagSets)
[14:05:33] <arnpro> hey kali I understand, I have been using .aggregate() with a $group clause, just fine, however I would like to limit the group limit count, to n documents.
[14:12:17] <arnpro> kali, does this helps http://pastebin.com/hPMw9E62 ? how can I re-write that into a mapreduce ?
[14:31:43] <bean|work> arnpro: you may want to give this a read: http://docs.mongodb.org/manual/applications/map-reduce/ and see if you can apply it to your current problem.
[14:37:45] <pquery> can anyone here help me with a sharding issue? I'm having trouble with an initial setup especially with connecting to the shards from the mongos intstance
[14:40:36] <Guest32592> hello, I want to find a document in an array in an array: inventory.squads[..].units[..]._id == mysearchid
[14:41:00] <Guest32592> I've used $all to search values in a single array level, but I need to search deeper now
[14:46:21] <Guest32592> If the second, then I'm using mongoose, but I'd like to know how to do this in MongoDB directly
[14:51:38] <arnpro> I think I started with the right foot, however the mapreduce is throwing an error: http://pastebin.com/e9NBgYR5 may someone tell me what I'm missing?
[14:56:45] <Guest32592> hmm, thanks NodeX I think. Turns out simply searching for { 'squads.units._id':mysearchid} works
[14:59:33] <bean|work> pquery: I set up mongos in front of a replicaset just fine last week. but i don't have actual shards, just a single replicaset
[15:00:45] <pquery> I'm running into the issue of the mongos not being able to connect to the shards
[15:01:34] <pquery> I've gotten past all the other issues and can get into each of the replica instances from the mongoS instance
[15:01:45] <bean|work> yes, i saw your previous message. I don't think I can help. How are you setting up the shard
[15:04:15] <pquery> all instances are running on 27017 except the config server which is running on 27019
[15:05:19] <pquery> I've tried restarting the mongoS instance a few times, and I've tried running the db.adminCommand( { flushRouterConfig : 1 } )
[15:13:10] <bean|work> ok, and then the command you're running for creating the shard is sh.addShard("rs_cent_shard/shardA:27017") ?
[15:22:01] <JoeyJoeJo> When I add .explain() to my find() queries, does "millis":4 mean it took 4 milliseconds to complete the query?
[15:22:23] <pquery> so I added 1 of the replica set, and then I tried to add another instance of the replica set and that failed...
[15:22:43] <pquery> does that mean that it has added all of the nodes of the replica set to the shard?
[15:32:29] <arnpro> hey bean|work I got this so far: http://pastebin.com/485Wcu7A however the reduce isn't working, it's storing everything. I need a way to debug those functions, is there an easy way for that?
[15:33:11] <bean|work> arnpro: I'm a sysadmin, not a developer, I was just simply pointing you to the documentatipn
[15:43:44] <arnpro> MatheusOl: I'd like to group all the docs belonging to and user defined by id_str with a limit of 10 per user, no more. so I'm trying to make a mapreduce function
[15:45:15] <arnpro> from what I have read, there is no $limit for a $group method, so I need to emulate it
[15:47:45] <arnpro> Im sorry MatheusOl I am not following you, I tried with agrgegate() just like http://pastebin.com/hPMw9E62 shows and didn't let me put a limit in there
[16:15:43] <arnpro> I thought about having a master array that counts from 0 to 10, if the count of user is 10 ignore it, something like that
[16:18:01] <arnpro> so that's the first step done, what would be the 2nd? "ignore the document if the count is over 10"
[16:26:17] <arnpro> there is no field, just count the docs the user has, if more than limit drop all the next ones
[16:47:36] <smith__> while trying to run mongo I got the following error: terminate called after throwing an instance of 'std::runtime_error' what(): locale::facet::_S_create_c_locale name not valid
[16:48:15] <smith__> I've tryed to re-gen locale by locale-gen en_US.UTF-8 Generating locales... en_US.UTF-8... done but the problem persists
[16:52:10] <rideh> how do i change datatype for a set of data from into to boolean? I imported a series of data and of course it didn't know how to interpret it.
[16:52:34] <MatheusOl> arnpro: If your reduce function uses tweets as an array, it may receive an array too
[16:52:50] <MatheusOl> arnpro: In some cases, MongoDB may send a result of a reduce function to another reduce function
[16:53:25] <MatheusOl> The second parameter of emit and the return of reduce must have the same schema
[16:54:26] <arnpro> alright, I see what you mean, so just reverse the order of params, and good to go, right?
[16:54:27] <kali> the second argument of the emit calls, the return value of the reduce MUST have the same form, and the second arg of the reduce calls will be an array of these same things
[17:00:00] <arnpro> MatheusOl: I see what you mean, however the new code is only storing 10 fields of each tweet :P
[17:05:38] <owen1> is it possible for a client to access the nearest mongo host in a replicaset (read freference = near) in version 2.0.1?
[17:12:18] <joec> I have a question on mongostat. Under the update column, is that reporting the number of update commands issues or the number of documents affected by an update during that time period?
[17:30:56] <arnpro> I got an starter question :p, how can I extract all the tweets from documents like these: http://pastebin.com/4zau1Zu1 and store them all together?
[17:45:23] <br-> hi. what does this mean? inserting a trivial document: pymongo.errors.OperationFailure: _a != -1
[19:05:40] <saml> i need to write a script that imports some data to mongodb.. and i was not sure about structure
[19:15:45] <arnpro> kali, how can I handle a document that does not necesarily fits into the reduce function of a mapReduce procedure?
[19:30:48] <arnpro> kali, The schema between the 1 tweet users and 1+ tweets users, is different, because the next step is to extract them all from there and put them into a collection
[19:31:35] <kali> arnpro: yeah, this is what i meant earlier: the emitted value must match the reduced result
[19:32:09] <arnpro> kali: I think it was worked out, tho? am I still missing something in the map/reduce functions ?
[19:33:52] <owen1> how to rename a key in a collection? this does nothing: db.devices_ypmobile.update( {}, { $rename: { 'udid': 'deviceId'} }, {multi: true} )
[19:40:20] <MatheusOl> notice that on emit you are sending ONE tweet, and on reduce you are returning one ARRAY
[19:40:35] <arnpro> here's the expected out, http://pastebin.com/kzF1Et2C, so even 1 tweet users, should be shown as array
[19:40:55] <owen1> kali: where does it say what the true and false means? http://docs.mongodb.org/manual/applications/update/
[19:48:04] <kali> owen1: http://docs.mongodb.org/manual/reference/method/db.collection.update/#db.collection.update
[19:48:20] <owen1> kali: is there a way to access old versions of the docs? i use mongo 2.0.1 and don't even know how to make my secondaries readable and make my app access the 'nearst' mongo host.
[19:49:19] <kali> owen1: not that i know of. TBH, i'd be glad if we had a well working doc for the current version. this documentation migration is a PITA and it takes forever
[19:50:25] <kali> owen1: also, mongodb is still very green. you really want to bump version (at least to 2.0.8) and consider bumping to 2.2.2
[19:50:33] <arnpro> kali can I modify the _ids when using $unwind? because when trying to move the tweets into 1 single collection the user_ids get duplicated
[19:56:15] <saml> json document is pretty large to look at in mongo shell. how do you look at json? can mongo shell fold some items.. etc?
[20:23:23] <JoeyJoeJo> What's a good way to convert a pymongo cursor into json? Right now I just iterate over the cursor and pass a dict to json.dumps. Is there a faster way?
[20:28:24] <pquery> I'm a fan of mongo, but mongoimport seems a bit slow, is there a reason that 50K documents would take over 12 minutes to import?
[20:28:34] <arnpro> kali: how can I workaround it? Itd be pretty good to insert each tweet the reduce function finds, but what about the unique ones?
[20:29:31] <kali> arnpro: and don't forget that mongo can call it several times for the same key to reduce in stages
[20:31:00] <arnpro> kali so im pretty much lost again hehe, my idea was to extract the tweets nested array from the mapreduce result and store then in a collection. as I have searched, it returns the whole document where your keyword was found, not just a part of it
[20:33:11] <pquery> CPRIVMSG kail mongodb :not that big, like 2 lines a piece I didn't count the bytes because one of the data team made them but I opened up the json and looked and they're small
[20:33:56] <saml> db.foo.findOne({"x": 1}) returns somethign like {"id": __ObjectId( ...)} which isn't json
[20:34:46] <arnpro> kali: I need to get the reduced results and the not-reduced ones, and store them apart
[20:37:52] <saml> mongo listings --eval 'printjson(db.Listing.findOne({"url": "ca-va"}))' > some.json
[20:38:54] <arnpro> kali: you mentioned something about generating _ids when using $unwind, how can I achieve that?
[21:58:30] <dan_> Hi! When I'm am in console in one of my mongodb shards, running db.stats() against all non-empty db's reports 26 indexes. But the MMS index count graph (DB Stats->Indexes) for the same shard reports 42 indexes. is there any way to reconcile the difference between these values? Per the deployed code, I expect 26 indexes.
[22:04:47] <dan_> maybe MMS is intrumenting mongo-internal system indexes that wouldn't be inventoried by db.stats() ?