Log file Viewer
#mongodb logs for Thursday the 14th of March, 2013
[00:01:16] <sinisa> http://docs.mongodb.org/manual/applications/update//#Updating-The%24positionaloperator
[02:21:36] <SecretAgentX9> I have a question about Mongoose schema's. Is this an appropriate place to ask ?
[07:56:22] <abhinav> Hi guys, QQ. I am using mongoid alias like "field :n, :as => :name, :type => String"
[11:46:00] <SirHirzu> I have a very fundamental problem, when creating the MongoClient instance ( val mongoClient = new MongoClient() ) Ii complains about parameters, what kind of parameters do I need there?
[13:23:54] <Nodex> http://images61.fotki.com/v545/photos/9/900497/5990888/whypeopleseemtohavefreetime-vi.png
[13:48:59] <SrPx> Hello. Does mongodb have incremental views similar to couchdb? Like, indexing mapreduce operations (if I understood properly)
[13:50:17] <SrPx> kali: so pardon but if I get it, mongodb is more similar to mysql with queries, but couchdb is faster if you prefer using mapreduce operations? Is that it?
[13:56:05] <kali> SrPx: this is kinda right. the thing is... i think couchdb introduced a lot of confusion by calling "map/reduce" what is actually materialized views
[13:56:45] <kali> SrPx: but we can safely say that mongodb has nothing that look remotely like materialized vied
[13:57:01] <kali> SrPx: and mongodb map/reduce is batch oriented. you can not use it for real time queries
[13:57:54] <kali> SrPx: you may want to check mongodb aggregation framework, it allows to solve quite nicely some of the problems that you can address with materialized views
[13:58:54] <SrPx> kali: I see! I've read some about it - it's the part that is kinda like SQL, right? Is there any reason it's faster than mapreduce?
[14:01:22] <SrPx> kali: also, I'm just learning about the view model. It looks really interesting. Say I'm storing 1k "players". I have a page that displays the ranking of players. I've been thinking about it... even if you dont know about it... Isn't the building of that page just a mapreduce of the list of players?
[14:02:39] <SrPx> kali: that is, first you select the interesting data (the player level and name)... then you reduce it (sort and create html rows)...
[14:05:22] <kali> SrPx: it's not a problem, as far as i'm concerned, i just want to make sure we're not confusing materialized views (a db feature) with the "view" of the MVC models (a UI framework abstraction)
[14:06:15] <SrPx> kali: I probably am, I don't know anything about MVC and am just learning about materialized views, but I made the association, they seemed like the same concept pretty much
[14:07:55] <kali> SrPx: but in most circumstances, you need some logic between what your database provides and what your user will manipulates in the UI
[14:09:52] <SrPx> kali: yes and that's the controller part of the "mvc" , which would be, eh... mapreducing the data ... am I completely mistaken as I feel I am? u.u
[14:11:43] <kali> SrPx: controller receive commands from the user, interacts with the model (which itself interacts with the DB) then prepare a view to be rendered, or alter what is displayed in a view, the view present stuff to the user
[14:12:21] <kali> SrPx: that's the theory. it works well with 1980's apps, where you have keyboard and a screen :)
[14:12:23] <SrPx> kali: I get it but in pratice whats the controller? The backend program? The server?
[14:12:49] <marcules> hmm I guess I could also ask here: Is there a way to not only group by a single character but a range of characters? So for example I have a ranged group from a to d and want all my users whose Last Name starts with a, b, c, d in a group returned?
[14:14:11] <kali> SrPx: the MVC terminology will often exists at several levels: for examples, spring or ruby on rails use it to organize a web app
[14:15:00] <kali> SrPx: but you can also use it inside a "javascript app" that get pushed to a browser for heavy ajax applications
[14:21:22] <kali> marcules: you're basically slurping your user collection in node.js space. i can't help you there
[14:22:38] <kali> medberry: and if you're a beginner, i strongly urge you to start something else than node.js
[14:22:52] <gazarsgo> ok so the mongodb cloudformation template starts off with a blank amazon AMI…would have been good to know :D
[14:45:34] <marcules> hmm I don't think node.js is the problem - rather this is my first project with mongodb
[15:06:29] <Derick> motiooon: having 4 doesn't gain you anything over 3 (but it doesn't hurt either)
[15:44:49] <capnkooc> hi guys, can someone please point me to the right direction on how to filter an array of objectIds with the timestamp store on the id?
[15:46:26] <capnkooc> I now how to do it using the object id of the document and filter the collection or a collection or objects but I really feel a little lost trying to do the same thing but with the objectIds inside an array
[16:11:49] <failshell> hello. does anyone have a backup tool/script for sharded clusters to recommend?
[16:46:24] <fommil> in the SQL mapping chart example, http://docs.mongodb.org/manual/reference/sql-aggregation-comparison/ what would be the mongo aggregation to get all the "item" entries as rows?
[18:00:27] <simon_> How can I verify that my three configservers know of each other and have the same data?
[18:16:23] <joe_p> simon_: to see if you config servers have same data : use config; db.runCommand({dbHash:1})
[18:27:50] <Rakko_> hi there.. I am taking my first steps to mongodb and I was wondering if you gurus know about sharding.. I know that sharding key cannot be changed "on the fly" but is it possible to kind of reconfigure the sharding like shard A contains all the records between a: 0-100, shard B contains a: 100-200, shard C contains a: 200-300 so that after my changes shard A contains all record where a: 0-250 and shard B contains a: 250-300
[19:01:39] <cookiem0nster> Trying to update mongo PHP driver to latest version, have been googling for a while and can't figure out the problem, can anyone help me out?
[19:11:16] <diminoten> what's a good way to estimate an upper end of capacity on a single mongodb instance
[19:11:42] <diminoten> like, I know the size of a single document, and I know that document will trigger other documents to be created, so the aggregate size of the created documents as well as the base document
[19:12:01] <diminoten> would constitute one object footprint, then divide the total available ram by this footprint value
[19:12:15] <diminoten> to get a sort of ballpark, order-of-magnitude value for what amount of data the app can handle before mongo performance tanks
[20:06:37] <varunv> Hey, I am trying to optimize indexes for a collection, where I mostly only access data upserted over the last hour or so. The data format is like: {_id: ObjectId( ... ), url : "something", clicks: {ref1: <count>, ref2: <count> ..., } }.
[20:07:16] <varunv> {_id: ObjectId( ... ), time: ISODate( .. ), url : "something", clicks: {ref1: <count>, ref2: <count> ..., } }
[20:08:25] <varunv> We retrieve the whole documents according to url and time. We have indexes on it, but it's too big to hold all of it in memory across all of the different datasets.
[20:21:27] <sinisa> in aggregation framework should one use $project{field:0, field:1} if wants only field 1 to be in stream?
[20:53:59] <ins0mnia> sinisa: { a: "b", arr:[{a:1}, {a:2}, {a:3}] } <- trying to pull for example {a:1} and {a:2}, anyway will try to pass an array of elements to pull see if it works
[21:02:41] <therealkasey> is it possible/safe to increase the size of an existing production capped collection?
[21:07:22] <therealkasey> is that a no sinisa? sure i could make a new collection, push the producers over to it, then wait until the consumers finished the old one, then move them over. but if it's possible to increase the size in place that would obviously be preferrable.
[21:12:25] <therealkasey> the docs do not specify one way or the other. you can convert an existing collection to capped, which gave me a glimmer of hope. hoped someone in here had experience or knowledge of internals.
[21:13:55] <sinisa> im looking for answer about aggregation .. also its not clear in the docs, should I use $project:field:1 to get ONLY field 1 in the stream
[22:01:01] <AAA_awright> Where am I supposed to go for information on the wire protocol and client development?
[22:07:54] <hjeong> not sure what the etiquette is here, but can I just start posting mongodb questions here?
[22:11:54] <nDuff> hjeong: ...trying to get people to volunteer to help before they've had a chance to hear your questions is poor form, so opening with questions first is generally much preferred. (And yes, thank you for asking).
[22:30:44] <hjeong> I have Students collection. Students has a list of Meetings (embedded), ordered by the field Meeting.time. And I want to do a query to get all students whose last meeting time is before 9pm. I expected {"meetings.-1.time"} to get the last meeting in meetings array, but it didn't seem to work. How can I query on the last element in the array?