Log file Viewer

[00:01:23] <dstorrs> np

[05:29:27] <LuckyAlgorithm> Hi all....

[05:50:27] <AndroidLoverInSF> when i do a find in mongo, it shows has more, how to see more?

[05:51:19] <mids> type 'it'

[05:53:06] <AndroidLoverInSF> mids, thanks! what does it stand for?

[05:54:00] <mids> iterate

[05:54:14] <mids> or itwasgoodIwantmore

[05:55:05] <ranman> mids: I loled

[12:06:35] <jammanbo> Are compound mongo indexes like MySQL in that they can be used in queries for any left prefix of the set of indexed fields, or no?

[14:27:53] <macUzer> how can I make mongo treat this { "StartTime" : "/Date(1340380800000-0400)/" } as timestamp?

[14:32:51] <deoxxa> macUzer: by making it into a date

[14:39:09] <macUzer> deoxxa: how? { $type : 17 } shows up as false. Is there a conversion function?

[14:43:56] <deoxxa> macUzer: { StartTime: Date(1340380800000-0400) } would do it

[14:44:12] <deoxxa> (as long as that's actually the date you want...)

[14:44:21] <deoxxa> (it's probably not - 0400 looks like a timezone)

[14:45:05] <deoxxa> depends on how you're inserting it as well. you're probably not just using the mongo shell, so you'd want to consult the manual for whatever driver you're using.

[14:50:10] <macUzer> yes, -400 is timezone. The rest is standard unix epoch time in seconds. I am indeed using the mongo shell, pulling JSON objects from a webservice and inserting directly using mongoimport.

[17:44:12] <Kryten001> Hi, is it possible to automatically remove the value field from a mapreduce output ? and get the data inlined at the root of the object ?

[17:49:52] <Gueste23131> nope

[19:22:14] <dstorrs> hey all. I have a collection of stats related to youtube videos. stats are added once a day, but we do not necessarily get every video every day. I would like to a find() condition that says "give me a cursor to the most recent stat for every video belonging to user X" and I'm having trouble. help?

[19:22:45] <dstorrs> docs look like this => { u : 'user_name', h : $harvested_epoch, .... }

[19:23:43] <dstorrs> it's easy to say "all of BOB's videos from time T". it's the "most recent avalable" part that is getting me.

[19:26:40] <dstorrs> if I do this => find({ u : 'BOB' }).sort({h : -1}) I get almost what I want, except I'll need to do duplicate elimination on client side, and I'll need to know how many videos there are to start with so I know when to stop

[19:32:44] <McNulty_> dstorrs You want .limit() on the end there

[19:34:42] <dstorrs> that would limit the entire thing though, wouldn't it?

[19:35:13] <McNulty_> oh, yes, sorry

[19:37:54] <dstorrs> oh...hm.

[19:39:18] <McNulty_> Well you could do a Map/Reduce but that might not be as performant as you want

[19:39:21] <dstorrs> actually, for this purpose it may not matter. If I only work off the latest harvest, and then I save what I end up with, it would update the ones I find and not the ones I don't

[19:39:40] <McNulty_> Frankly it'd be easier to tag the 'latest' stat each time

[19:40:01] <dstorrs> ?

[19:40:48] <dstorrs> how would that work? I've got 10M videos overall, spread unevenly across 30k channels. Each channels has 0 - 50,000 videos. We harvest every video every day, and bulk insert their data into the stats collection.

[19:40:57] <dstorrs> *their point-in-time data

[19:41:32] <dstorrs> how do I keep a 'this is the latest' tag on each video, keep it up to date, and do it performantly?

[23:18:23] <multiHYP> night all