pmxbot IRC Log Viewer

[03:08:07] <Astraea5> Hi. I'm trying to find N of each type of something in a collection. Like a collection of objects indexed by color, and I want two objects of each color (but I don't know what colors exist beforehand).

[03:08:38] <Astraea5> Can anyone point me in the right direction to make that sort of query?

[09:37:12] <jet> Is there a way to set a long value into a document inside the console?

[09:58:38] <cheeser> NumberLong(4) iirc

[10:00:48] <jet> thanks

[11:49:45] <will_p> Hey. I have created a user with the 'readWrite' role inside a database, but when I authenticate with them I can't insert. I get this error (using the mongodb ruby client): 16544: not authorized for insert on my_db.my_collection

[11:49:57] <will_p> Anyone have any idea what the problem could be?

[11:51:43] <will_p> It seems to work fine via the mongo client using the same user :/

[11:53:21] <joannac> You sure you authenticated correctly?

[11:58:47] <will_p> I think so. Printing out client.auths shows the expected authentication

[12:17:56] <will_p> Figured it out. Missed the read_only argument and so the options hash with roles: ['readWrite'] was truthy and therefore made it read only. Good times.

[12:45:51] <andrepadez> Hi everyone! can someone help me with a design problem i'm having? I'm building a blog with mongodb. I am having trouble figuring out what would the best way to find the next and previous post, counting from the one i'm reading now

[12:58:57] <Nodex> andrepadez : $gt : {ObjectId()} -> limit 1 sort by _id asc

[12:59:05] <Nodex> and the reverse for previous post

[13:05:36] <andrepadez> Nodex: that is brilliant. I had no idea that would work

[13:05:40] <andrepadez> :) thanks

[13:06:30] <andrepadez> I'm gonna read on ObjectIds, i didn't know they could be seen as sequential

[13:07:48] <rcp> hi anyone up for paid mongodb project? or know where I can find someone

[13:08:12] <Nodex> andrepadez : they're timestamp based so you also get a free sort() on them as they're already indexed ;)

[13:12:06] <andrepadez> Nodex: :) once again, thank you

[13:24:09] <Nodex> ;)

[13:53:03] <asker> hey guys, i have an array in a doc and i need my findOne to return only certain array elements matching my criteria. i don't get to filter the output

[13:53:45] <Nodex> look at the $ oprtator

[13:53:53] <Nodex> +projection

[13:57:21] <asker> Nodex: the $ limits the array output to only the first matching element, i need to return multiple elements that match my needs

[13:57:36] <asker> is it possible somehow?

[13:58:08] <Nodex> try it with $elemMatch

[13:58:13] <Nodex> you might get lucky ;)

[14:02:08] <asker> Nodex: http://pastebin.com/E5QdSvNy stil returns only the first element

[14:03:36] <Nodex> then it probably can't be done

[14:03:52] <Nodex> can't you just filter the array in php?

[14:17:39] <asker> it might have a lot of elements which i don't need in the query, that's why i didn't want to filter the array in php

[14:17:57] <asker> -> more traffic & slower

[14:19:04] <Nodex> I don't think you're going to have a choice

[15:13:52] <zexcriz> is learning mongodb more eady than any DBMS out there ?

[15:15:13] <Nodex> eh?

[15:15:43] <zexcriz> i don't have any of the DBMS like mysql, or postgres

[15:16:09] <Nodex> what does "Don't have any" mean?

[15:16:39] <zexcriz> Nodex, i don't know mysql

[15:16:45] <zexcriz> i am new to database

[15:17:12] <asker> quite the same diffuculty i would say sql vs nosql

[15:17:22] <Nodex> Personaly I would say it's easier to pick up when you don't know all the limiting factors of RDMBS based systems

[15:17:39] <Nodex> being that a lot of NoSQL type concepts are the polar opposite

[15:19:06] <zexcriz> asker, ok

[15:19:23] <zexcriz> Nodex, good, any place where i can start learning about mongodb ?

[15:20:09] <Nodex> zexcriz : mongo do online courses for free

[15:20:29] <zexcriz> Nodex, any link to it/

[15:20:39] <Nodex> google ;)

[15:21:02] <Nodex> http://www.openpropertyexchange.co.uk/ <--- second result

[15:24:56] <zexcriz> Nodex, thank you :)

[15:31:31] <Nodex> no probs, good luck

[15:32:36] <dcrosta> zexcriz see https://education.mongodb.com/ also

[15:33:26] <zexcriz> dcrosta, cool :D

[15:33:31] <asker> heyyo ;) do anybody knows why it concurrent sql db connection are so much more resource-expensive compared to mongodb???

[15:33:36] <Nodex> zexcriz : sorry, wrong paste

[15:33:58] <Nodex> silly clipboard haha

[15:35:09] <Nodex> asker : mongodb pools the connections

[15:35:18] <asker> what does that mean?

[15:36:40] <Nodex> http://docs.mongodb.org/meta-driver/latest/legacy/notes-on-pooling-for-mongodb-drivers/

[15:39:52] <asker> how does sql do it?

[15:41:19] <Nodex> I don't have a clue, at a guess it would probably hold the socket open for each request until explicitly closed which uses memory etc but that's just speculation

[15:43:25] <asker> well ok, need to checkout the diff. you get 'unlimited' mongodb connection for free hosting and like 50 concurrent mysql connection costs about 100€ on a dedicated mysql server

[15:45:57] <Nodex> why do you need persistent connections in mysql?

[15:51:44] <GHJGH> afternoon all, I'm doing my first mongodb upgrade, 2.4.4 to 2.4.8, any gotchas or tips?

[15:59:32] <asker> Nodex: i have several workers running... they need mysql connections

[16:00:40] <asker> i tried to wait for available connection if there are no slots free. connect -> do stuff -> disconnect but it won't work with ratio like 2:1 (workers:connections)

[16:13:42] <Nodex> asker : what server side language are you running?

[16:13:52] <asker> php

[16:14:01] <Nodex> are you running it as an apache module?

[16:14:17] <Nodex> or under fpm?

[16:14:37] <asker> ähm

[16:14:50] <Nodex> duno what that is

[16:16:17] <asker> its *thinking* well, i don't know anything about setting up servers myself. so im hosting it on a cloud provider (fortrabbit). they use scheduler, i guess its using the fpm

[16:17:38] <Nodex> fpm is not threaded it's a spawned process from a parent so I'm not sure how that gets handled in the connection. It's been a few years since I touched mysql in any sort of meaningful way

[16:18:57] <asker> well, im close to completing my mongodb adapter, so... no need to care about mysql anymore :D

[16:23:29] <Nodex> bonus ;

[16:23:30] <Nodex> ;)

[16:31:15] <zexcriz> Nodex, np :)

[17:13:49] <banditron> i need to change the datatype of a field on many documents in my db. if i'm running a mongo shell on the same machine as the master, it is better/worse to use db.eval() { ... forEach() } ?

[17:14:16] <banditron> vs db...find().forEach()

[17:14:28] <justanotherday> Hello, I need to insert a large sql dump file into mongodb, can anyone please help explain how to do this properly

[17:15:48] <cheeser> well, mongo can't read a sql dump. so you'll have to convert it to something that mongoimport can read.

[17:15:56] <cheeser> json, csv, or tsv

[17:16:35] <jyee> unless you're asking how to actually store the dump file as a file in mongo?

[17:18:11] <justanotherday> yea, I probably need to write a praser to convert the mysql data into json, cant seem to find one online

[17:19:25] <justanotherday> No, I dont wanna store as a file, want to convert and run queries on this file

[17:19:46] <justanotherday> but am not so sure how best to go about it

[17:19:53] <jyee> the sql dump is just one mysql talble?

[17:21:11] <justanotherday> I just one sql dump file

[17:21:19] <justanotherday> *have

[17:21:34] <jyee> but a dump file can have multiple tables.

[17:21:54] <justanotherday> Yea, I dont know the nature of the file

[17:22:21] <justanotherday> I was told it has 4 tables

[17:22:53] <cheeser> you'll want each table in a separate file

[17:22:53] <jyee> and how do the 4 tables relate (or maybe they do not)?

[17:23:20] <cheeser> yeah. you might bea ble to generate nested docs depending on the data.

[17:23:29] <cheeser> but it's probably not quite that simple

[17:23:59] <ron> cheeser: "be able"

[17:24:03] <jyee> not that simple and generally not a good idea to just throw a relational schema into mongo and expect any usefulness from it.

[17:24:20] <justanotherday> Yes, I guess there is a relationshop between somehow as I am supposed to perform a JOIN when I get it all set up ( i think mongodb does not support join) but alas I have to work on this

[17:24:34] <jyee> ugh

[17:24:38] <cheeser> ron: that's not how it's said here in yurop

[17:24:43] <cheeser> you cretin

[17:24:53] <ron> still in ireland?

[17:25:06] <cheeser> geneva. dublin was only a layover.

[17:25:11] <ron> aah

[17:25:14] <ron> merde!

[17:25:28] <ron> geneva is *expensive*

[17:25:31] <cheeser> so even though i speak 0 french, my thoughts have a french accent

[17:25:44] <ron> all you need to know is 'merde'

[17:25:54] <jyee> justanotherday: you should figure out what joins you need before hand, then write a quick script to do those queries, format the data and insert them into mongo.

[17:26:03] <cheeser> that's "thank you" right?

[17:26:14] <ron> cheeser: sure

[17:26:48] <cheeser> great. i'll be sure to use it at dinner tonight.

[17:27:05] <cheeser> what would I do without friends like you?

[17:27:25] <ron> I don't know.

[17:27:34] <ron> too bad I'm not in geneva now. could have met.

[17:28:06] <cheeser> i know! i'll be in dusseldorf friday. frankfurt. gothenburg. take your pick. :D

[17:28:10] <justanotherday> I kinda know what I am supposed to be, but how I am supposed to do it is the problem. Am not an IT person, reading biology, we usually use sql to store data but am supposed to be investigating mongodb and i have all this info, but totally lost on how to get it done

[17:28:38] <ron> not in europe anymore now that I'm out of a job

[17:33:55] <retran> europe is so rude to kick you out just b/c you dont have a job

[17:56:55] <rafaelhbarros> anyone with ndb experience?

[17:57:18] <rafaelhbarros> I know, "wrong" chat, but as someone with mongodb exp, I need some help migrating the ndb stuff.

[17:58:28] <jkitchen> I'm working on doing some offsite backups for our mongodb cluster and I'm wondering if it's save to put the dump directory inside the mongodb data directory?

[17:58:54] <jkitchen> ubuntu uses /var/lib/mongodb for data directory by default. I want to dump to, say, /var/lib/mongodb/backup-dump

[17:59:23] <jkitchen> reason being that /var/lib/mongodb is a different volume than /, and / is not very big

[17:59:36] <jkitchen> but /var/lib/mongodb is really damn big :)

[18:00:21] <Nodex> you can dump anywhere you like

[18:00:30] <Nodex> by default it goes in the current directory

[18:00:45] <jkitchen> right, but will mongo care if I dump into its data directory?

[18:00:53] <jkitchen> I know mongodump itself won't care.

[18:13:07] <bjori> no it doesn't

[18:13:34] <bjori> jkitchen: but for a backup system.. its not very productive to dump onto the same disk..

[18:14:45] <bjori> oh. I see. you already knew that :)

[18:16:49] <jkitchen> bjori: I have to dump it locally before shipping it offsite

[18:17:17] <jkitchen> I'd rather not loop through each db/collection and ship them individually

[18:17:43] <jkitchen> and I can't pipe the wholedamnthing because it doesn't support that right now

[20:15:48] <faranjessani> hey guys, is there a good reference for the different java frameworks out there to communicate with MongoDB w/o tight coupling in the models?

[20:16:26] <cheeser> what do you mean by tight coupling?

[20:16:40] <faranjessani> my objective is to avoid mongodb specific annotations in the models

[20:17:19] <cheeser> oh, well, that's going to be really hard to do.

[20:17:58] <faranjessani> hah awesome

[20:18:21] <faranjessani> is there no equivilant [Serializable] from c#?

[20:19:06] <cheeser> you could use eclipselink or hibernate obm if you don't mind *those* annotations. or spring-data. i prefer morphia for not so subtle bias reasons.

[20:19:19] <cheeser> well, you could use mongo-jack i think it's called.

[20:19:27] <cheeser> then you'd jsut have jackson annottions.

[20:20:03] <cheeser> (stupid hotel wifi messing with my typing and i just can't be bothered to fight it anymore :) )

[20:20:11] <faranjessani> awesome, let me look that up.

[20:20:11] <faranjessani> thanks

[20:33:45] <bjhaid> Hi all,

[20:33:45] <bjhaid> I am having issues with box query:

[20:33:45] <bjhaid> what is the difference between the two:

[20:33:45] <bjhaid> selector: {"location"=>{"$within"=>{"$box"=>[[-87.678096, 41.93223], [-87.667361, 41.961456]]}}}

[20:33:45] <bjhaid> selector: {"location"=>{"$within"=>{"$box"=>[[-87.667361, 41.961456], [-87.678096, 41.93223]]}}}

[20:34:01] <bjhaid> I expect that they should return the same values, but the don't

[20:35:08] <cheeser> order matters: http://docs.mongodb.org/manual/reference/operator/query/box/

[20:43:07] <bjhaid> cheeser: there is nothing on that page that shows order of the coordinates

[20:49:30] <cheeser> [ [ <bottom left coordinates> ] , [ <upper right coordinates> ] ] } } }

[20:59:15] <bjhaid> cheeser: if I want to get everything in the triangle how do I solve that

[20:59:23] <bjhaid> irrespective of the order

[21:00:00] <cheeser> well, you wouldn't use $box for starters.

[21:00:11] <cheeser> there's $polygon

[21:05:21] <bjhaid> cheeser: I have 2 points and I want to get coordinates within the 2 points

[21:06:42] <cheeser> 2 points make a line. or if you infer some lines a box. neither is a triangle.

[21:07:21] <bjhaid> okay

[21:07:26] <bjhaid> I get you

[21:07:47] <cheeser> bbaib

[21:08:59] <bjhaid> I get you, guess the name box confused me

[21:33:47] <natefinch> hopefully quick question - is there a database command that duplicates the functionality of rs.config()? I can't seem to find documentation on it. I'm trying to write some Go code using labix.com/v2/mgo that manipulates replica sets.

[21:39:16] <jkitchen> natefinch: function () { return db.getSisterDB("local").system.replset.findOne(); }

[21:39:33] <jkitchen> natefinch: if you run rs.conf without the parentheses that's what you get

[21:40:22] <natefinch> jkitchen: ahh, yeah, I see now it lists local.system.replset ... which makes sense, since it's just a document in the system

[22:01:10] <jrbasso> has someone had an issue of the moveChunk being aborted in a shard? I have the version 2.4.4 and it is showing in the log multiple times the moveChunk was aborted

[22:17:31] <jrbasso> I got the answer from https://groups.google.com/forum/#!topic/mongodb-user/6usJtSMtUiQ

[23:32:53] <testing22> hello, i've followed the Use Cases > Pre-Aggregated Reports page (http://docs.mongodb.org/ecosystem/use-cases/pre-aggregated-reports/) to develop my reporting system, however i'm running in to issues wrapping my head around timezones using this technique. from what i've gathered researching this, in order to allow reporting in different timezones, i would either have to aggregate in to supported/specific timezones, storing redundant d

[23:32:54] <testing22> for each or aggregate on a granularity no greater than an hour... thoughts ???

[23:34:53] <testing22> the problem is that with my data, if it were to aggregate on an hourly basis, it would be so segmented that the pre aggregation benefit wouldn't exist and i'm back to square one

[23:37:43] <aodhol> hello

[23:39:40] <aodhol> i'm trying to insert new documents into my database. The schema has several nodes, one of which is an array. When I insert I wish to have items which have the same compound key updated such that the array node will have an element added to it

[23:39:42] <aodhol> is this possible?

[23:54:17] <testing22> aodhol: check out the $addToSet operator

Log file Viewer

Help | Karma | Search:

#mongodb logs for Monday the 18th of November, 2013