Log file Viewer
#mongodb logs for Saturday the 25th of August, 2012
[01:59:20] <Xabster> Hi, I've tried this http://pastebin.com/TbieRMBN - I expected to get a resultset limited to 15 sorted results... and they kind of are that, but everytime i run the query I *can* get different results even though the data is not changed... any ideas?
[03:15:06] <MikeFair> meghan: Hi, kind of amazing that's there's over 300 users, and everyone stays quiet :)
[03:49:46] <Xabster> Hi, I've tried this http://pastebin.com/TbieRMBN - I expected to get a resultset limited to 15 sorted results... and they kind of are that, but everytime i run the query I *can* get different results even though the data is not changed... any ideas?
[05:24:17] <Xabster> i do a sort and then a limit but the results change even though the collection has the same data
[05:28:01] <IAD> Xabster: add "date_created" field into documents and fill it manually, when document be saved
[08:26:43] <JoeyJoeJo> What is the mongo command that will do the same thing as SELECT columnname FROM table?
[08:29:25] <JoeyJoeJo> No, I know the fields. I just want to return one specific field from every document
[08:31:00] <Xabster> is it correct to find(), sort(), and limit() in that order when you want to latest 15 entries? - the sorting is on a timestamp field
[08:35:03] <IAD> Xabster: the mongo driver will make query to database when you will try get first value from cursor
[08:48:45] <Xabster> i do those 3 lines to find a single entry ("youngest" entry in the 'datetime' field)
[08:54:56] <IAD> and use $cursor = $cursor->limit(1); before fetch value from cursor. if you want only one value
[09:00:59] <IAD> documentation: limit() is analogous to the LIMIT statement in MySQL: it specifies a maximum number of results to return. For best performance, use limit() whenever possible. Otherwise, the database may return more objects than are required for processing.
[09:01:01] <IAD> http://www.mongodb.org/display/DOCS/Advanced+Queries#AdvancedQueries-RetrievingaSubsetofFields
[09:04:24] <cubud> If my db is going to have billions of objects in a single collection would it be worthwhile me using the immutable ID (which is a string) and creating a hash of it, and then using multiple collections MyCollection_{hash}
[09:08:44] <cubud> Makes no sense to write a software solution to a problem I currently do not have and that is already solved by the DB :)
[09:26:35] <NodeX> off the top of their head does anyone know how to get a facebook page id (a fan page I suppose you would call it)
[09:32:54] <Xabster> well, it's about as relevant and proper as you asking for facebook IDs in a #mongodb channel on IRC
[14:55:09] <emocakes> I was only using SQL due to familiarity and I want to push myself to learn something new
[14:55:31] <emocakes> my models aren't very complicated, I would need a standard user profile document with some statistics on it
[14:55:49] <emocakes> I know if I ask if MongoDB is suitable for this, the answer will probably be yes
[15:22:30] <kali> I hate javascript, and i think node.js is generally not used for what it is intended
[16:35:37] <mikeywaites> hey hoping someone can help, mongo has completely died on my server, its not running and it just falls over everytime i try to restart it. can anyone suggest anything i can try to figure this out?
[19:23:29] <ppetermann> i seem todo something horribly wrong, i got a collection with 709068 documents, but an explain on a find shows me 2316051 nscanned, even with an index that should cover all fields, so i seem to be missing something
[19:49:04] <ppetermann> it seems the sort i'm doing is totally killing me, without the sort its all like i expect it, as soon as i add the sort, it goes downhill
[20:01:59] <kali> ppetermann: show us a getIndexes() on the collection, the query, and the explain()
[20:04:51] <ppetermann> http://nopaste.info/749b3f57fe.html document, http://nopaste.info/39c85c240f.html explain, http://nopaste.info/97027261f9.html the query
[20:08:29] <ppetermann> if i dont hint() it, its using one that has three more fields in it, which aren't used for this query
[20:09:49] <ppetermann> as i said it seems to be the sort that kills it. withou the sort the or can use two single field indexes and is fast
[20:09:58] <kali> you can either split it in two and have two indexes (attackers.corpId, killTime) and (victim.corpId), one for each query
[20:10:38] <ppetermann> thats not really helping me much, because then i have to resort in the application
[20:11:08] <kali> or you need to add an hybrid field (inConflict.corpId) contaning both ids of attackers and victim, then add an index on (inConflict.corpId, killTime)
[20:11:57] <ppetermann> it might be more than one attackers id, but yeah thats the worst-case solution i've come up with aswell, but i was hoping to find a better way
[20:12:24] <kali> if you don't understand why, i strongly advie to loog at the presentation there: http://www.mongodb.org/display/DOCS/Indexes#Indexes-Presentations
[20:13:43] <ppetermann> i wish there was a way that i could run the sort on the result of the find()
[20:14:16] <ppetermann> ah, those are reflecting the names of the API its pulling the data from and therefor out of my hand
[20:15:20] <kali> because, contrary to what you may be used too if you come from the SQL world, the field names are stored for each document
[20:16:24] <ppetermann> i'm aware of that aswell, but thats ok, since it improves the maintainability alot
[20:17:04] <ppetermann> especially having the names the same as from the api they are comming from enables guys who are shy about touching it (because of mongo) to at least semi-understand whats going on
[20:17:40] <ppetermann> and since its opensource the understanding of those touching it is quite important
[20:18:46] <ppetermann> i haven't really got why people are so scared of mongo, in most cases its quite simple in use - except if you run into indexing stuff like in this case
[20:18:51] <kali> but anyway, if you want to get this results efficiently, you need a involved: [ corp1, corp2, ... ]
[20:20:36] <kali> and SQL/oracle marketting and lobbying as made nearly synonyms the words "oracle" and "database"
[20:21:23] <ppetermann> well the famous "webscale" video where they make fun of it might be one of the reasons too
[20:23:01] <kali> i think the various blog entries explainging how mongodb hurt their platform, their data or their business is more harmful
[21:12:01] <Yiq> mongodb is not relational but is it suited for a task where you store blog posts by title and text? or basically if you need mysql or postgres you dont need mongo?
[21:23:42] <MikeFair_> SQL is really cumbersome to work with hierarchical data or data where the records don't have the same types of attributes
[21:24:41] <MikeFair_> But that doesn't mean that some aspects of your appliation wouldn't find a relational DB useful
[21:25:48] <MikeFair_> Synchronizing with other applications like mobile apps is another area where the NoSQL stuff seems to be doing really well
[23:58:36] <Yiq> ERROR: listen(): bind() failed errno:98 Address already in use for socket: 0.0.0.0:27017