Log file Viewer

[07:00:10] <Zelest> i wish to update a sub-document, if it exists.. otherwise add it to the set.. how can I do this in one query?

[07:08:50] <jelle> Zelest: find_and_update I assume

[07:08:57] <jelle> saw it somewhere in the mongodb docs

[07:09:03] <Zelest> OH!

[07:09:09] <Zelest> it's from HERE i recognize you :D

[07:09:09] <jelle> well python bindings that is

[07:09:20] <Zelest> :D

[07:09:28] <Zelest> i'll give it a go, cheers :)

[07:17:18] <jelle> Zelest: I guess you want something like {'foo.sub.1': 1'}, update: { "$set": { 'foo.sub.1': 1 } } ?

[07:21:08] <Zelest> solved it using two queries.. :P

[07:21:21] <Zelest> yeah, one for $set .. if the sub-document exists..

[07:21:40] <Zelest> if that returns 0 modified, do a $addToSet instead :)

[07:22:07] <jelle> ohh

[07:22:13] <jelle> I think you can do it with one :-)

[07:22:21] <Zelest> probably :)

[07:24:41] <jelle> I wonder if you can sort a subdocument on query

[07:27:35] <Zelest> i think that need some aggregation magic

[12:50:30] <SimpleName> after I addShard to cluster with mongos, now when I use enable Sharding(“database”), it show me need auth, so how to auth , use admin , db.auht() does not work

[12:56:43] <SimpleName> not authorized on admin to execute command { enableSharding

[13:16:47] <SimpleName> ok, I solve it

[13:16:56] <SimpleName> need create a new user on mongos

[14:10:50] <suicidel> hi

[14:11:20] <suicidel> I am trying to install mongo following the install guide. After importing the key and adding the repo, I get unable to locate mongodb-org when I install....

[14:37:21] <AlmightyOatmeal> wow, so much help here on IRC...

[14:37:29] <AlmightyOatmeal> don't everyone jump in at once and help people out here

[14:37:52] <AlmightyOatmeal> it really turns me off of MongoDB and makes me want to find another solution where people actually give a damn.

[14:41:30] <cheeser> 1. it's a holiday in the US. 2. there are also mailing lists. 3. people *do* give a damn.

[14:41:58] <cheeser> if *you* have an answer for suicidel, then by all means answer.

[14:51:59] <AlmightyOatmeal> cheeser: i've been asking a relatively simple question for days but i've been looking at the number of questions coming in vs the number of answers being given (or even advice)

[14:52:18] <AlmightyOatmeal> so it is rather disappointing.

[14:55:17] <cheeser> this isn't the main support forum. that's the mailing list. there are some of us here who do what we can but we're not always here and we don't always know the answer (or have the time to follow through)

[15:13:59] <adrian_lc> hi, is an "unordered bulk operation" with only updates roughly equivalent in performance to updateMany ?

[18:03:14] <AlmightyOatmeal> can i get a value from a document based on a string within the document while doing an aggregation?

[18:05:06] <cheeser> "$someField"

[18:05:59] <BadApe> is there a ui to help me design map reduce functions like couchdb futon?

[18:06:15] <AlmightyOatmeal> cheeser: if the value is stored in $_source._properties then i can't do $_source._source.properties

[18:06:42] <cheeser> why not just $_source.properties?

[18:07:36] <AlmightyOatmeal> cheeser: because $_source._properties is the string of what i want to get from $_source.

[18:08:30] <cheeser> gtg

[18:08:34] <AlmightyOatmeal> for example, {'_source': {'tacocat': true, '_properties': ['tacocat']}

[18:08:59] <AlmightyOatmeal> so i do an $unwind and $_source._sf_properties becomes the string in the array

[18:11:25] <AlmightyOatmeal> so thinking of it like a python dictionary, i'm trying to do: _source[_source._properties] (with _source._properties being the unwind value

[18:11:28] <AlmightyOatmeal> )

[22:59:48] <SET001> I'm building a game where each account can have a bunch of characters. Should I store those characters as part of accounts document or should they be stored in separate document?

[23:00:38] <SET001> I'm very new in noSQL and it's kinda hard to make used to it after relational DBs

[23:10:35] <cheeser> i would with separate docs

[23:10:43] <cheeser> would *go* with...

[23:11:19] <earendel> hi.. does mongodb use some kind of compression internally? the reason i ask is following: i have a collection storing objects with fieldnames like {"server": "irc.freenode.com", "channel":"mongodb", "nick":"earendel", "message":"hello world", "time":"blah"} ..now with thousands of entries my db grows kinda large. and i asked myself it would make any sense to use shorter fieldnames to save on space.. so for the

[23:11:20] <earendel> above like {"s": "irc.freenode.com", "c":"mongodb", "n":"earendel", "m":"hello world", "t":"blah"}

[23:13:44] <earendel> i guess for hence my entries are all of the same structure a documentdb like mongo wasn't the best choice anyway right?

[23:14:02] <cheeser> i use mongodb for an irc bot, too, fwiw.

[23:14:21] <cheeser> and yes, you should use shorter field names if disk space is a concern

[23:14:37] <earendel> hm. ic.

[23:14:43] <earendel> tanks

[23:15:46] <blizzzard> Hey how would you setup a schema to support a User rating/favorite system?

[23:15:50] <earendel> maybe one more.. i know relations are not the domain of mongodb, but is there any way to have relations handled on db level even in mongo?

[23:16:25] <blizzzard> The only kind of databases I've set up are basic string based ones so I have absolutely no idea how to approach this.

[23:16:51] <blizzzard> I've thought of something like favorites: [{ type : ObjectId, ref: 'User' }]

[23:17:09] <earendel> i have like 3 servers..but every log entry contains the full host name.. which is annoying (in puncto space)

[23:17:21] <blizzzard> such that favorites would be a collection of ID's

[23:17:40] <blizzzard> is that the right direction? how do I achieve a rating system?

[23:19:26] <earendel> one solution i can think of is to have another collection with an enumeration of servers, and storing their ids in the logs, and do the match in the application..

[23:24:08] <earendel2> i could also use a collection for each host.. if anyone has ideas or suggestions on that.. please ping me. thanks.

[23:24:58] <blizzzard> It's dark in here

[23:29:29] <earendel> blizzard: usually u would go for a m:n table .. like Object (PK:ObjectId) <-> UserRatedObject (PK:RatingId,FK:ObjectId, FK:UserId, Rating) <-> User (PK:UserId)

[23:29:51] <earendel> UserRatedObject would be the m:n table

[23:32:07] <earendel> that would make 3 tables in total.. that's the relational approach