Log file Viewer
#mongodb logs for Saturday the 18th of January, 2014
[04:25:37] <TehShrike|Work> I'm saving documents with an array/set in it - how would I go about finding the document with the largest number of items in that list, or even just find the largest count of items in a set?
[07:21:35] <lgp171188> I am trying to insert a python dictionary into a mongodb collection. The python dictionary is obtained by running the serverStatus command from pymongo. But when I try to insert the data, I get "bson.errors.InvalidDocument: key '.' must not contain '.'" What could be the issue here? I don't see any keys having a . in their names.
[07:36:53] <regreddit> why does this query work: users.findOne({ user_id: {$regex: userid, $options:'i'}}, but this does not: users.findOne({ user_id: {$regex: userid, $options:'i'}, field2:'bar'}?
[07:41:53] <mrpoundsign> db.pixels.findOne({color: {$regex: /#/}, x: 1}).x resulted in NumberLong(1)
[07:42:33] <regreddit> mrpoundsign, can you try something for me? make the value for regex a variable, vs a literal
[07:45:32] <mrpoundsign> db.pixels.findOne({color: {$regex: "#", $options: 'i'}, x: 1}).x -- same result
[07:47:16] <regreddit> i know expressing regex syntax as strings, in variables can be rticky in javascript
[07:49:26] <mrpoundsign> http://stackoverflow.com/questions/10728043/mongo-query-with-regex-in-node-js-operating-on-a-variable ?
[07:50:15] <regreddit> just figured it out - field2 is a number in the db, so when i build the variable, I need to *1 it
[07:58:46] <mrpoundsign> wrote http://www.pixelparty.us/ -- every pixel is in Mongodb. It runs in < 40mb RAM, all over websockets. I'm having a blast :)
[07:59:42] <mrpoundsign> like, I am generating the export images on-the-fly haha. It's not optimized at all.
[08:23:48] <lgp171188> I ran db.serverStatus() command on mongodb and in the data returned, there is a key '.' in the value of 'locks' key. Does mongodb allow . in the key name?
[13:10:58] <strangewildthing> is mongodb the right choice to use for processing 'limit orders' in a 'central limit order book'?
[13:16:20] <kali> what you descrive seems just right the only thing in the industry i would stick to a sql database
[14:37:24] <d0x> Hi, i've wrote a aggregation query (with unwind and one projection) to get a flat json that i need to export as CSV for my boss. But the result is larger then 16MB. Is that an identicator that mongo db is not the best database for our usecase or should i use mapreduce or smth. like that?
[14:44:19] <d0x> hm, i think map reduce is not the beneficial for me. Because i like to transform the whole collection
[14:47:22] <ppetermann> i'm not sure what you are doing there, neither map/reduce nor the aggregation framework are really for transformation..
[14:51:11] <kali> well, you can map/reduce to get the result you want in a new collection (who cares if there is no reduce) then dump the resulting collection with mongodump
[17:52:17] <lampe2> i have a collection items every object in that collectoin has a entry "category" this category is a Array of strings ["cat1","cat2"] . What is the best way to find all strings and return a object/array with only that strings? Example: item1.category = ["cat1","cat2"] and item2.category = ["cat2","cat3"] and i want catgeory = ["cat1","cat2","cat3"]
[21:01:09] <cm0s> How do you usually translate a many to many relation in Mongodb? Do your create two documents with a field containing an array of ObjectId to keep the references? Or do you keep the reference on only one document?
[21:16:07] <cm0s> I have a "relational db" background and changing the way to think about it is sometime difficult ;)
[21:19:21] <kali> instead of thinking of the data in a static way, you need to think about the access
[21:19:22] <kali> try to picture what your application will require, and what data layout can provide the data in an efficient way
[21:21:07] <kali> shingara would say "one http hit, one single mongo document read". (i think it is extreme)
[21:24:23] <cm0s> When you need to modify one element you will need to modify it everywhere it has been embedded.
[21:25:25] <cm0s> But it seems to be a good idea if you rarely modify the document. Accessing the document will be extremely fast!
[21:27:54] <kali> cm0s: just one thing: getting data out of mongodb tends to be way faster than a sql engine
[21:28:42] <kali> cm0s: so making two queries (one to get bookmarks, then on to get category) with no denormalisation could get you quite a long way
[21:29:35] <kali> also, categories are usually a quite small dataset, that will fit in mongodb ram (making access fast), or even in your application ram, making them negligible
[21:34:44] <cm0s> Do you think if we create a totaly normalized schema on mongodb the perf for getting data should be pretty similar (or even worse) than on sql?
[21:36:53] <kali> no, i actually think if you craft the queries sensibly (like i said, read a page of bookmarks, then read all categories, so two queries), you can get better performance than relational databases
[21:50:53] <cm0s> But if you make two separate queries you'll then have to "reassable" the to documents. If, for example, you need a json document with an array of categories which all have their associated bookmarks embedded in the same document.
[23:14:43] <swak> http://pastebin.com/eZFnp6YT Any ideas on having the "target.update({b" use the key parameter instead of something set like 'b'? This is with javascript.