[04:47:31] <SrPx> Hello guys, why are you using mongodb over other nosql dbs? That is, what you like more about i?
[04:48:01] <Derick> I can only give a biased answer :)
[09:07:01] <Hounddog> trying to do a pagination but a little stuck here somehow... i am sorting descending, and getting the first batch with limit 10 after this i saving the last id i get from that call and want to get the next 10 items
[09:07:12] <Hounddog> i can not use skip as data is beeing continuously added
[09:07:33] <Hounddog> so i have to go something like where id is smaller then saved id
[09:51:54] <kali> it shouldn't, update to a single document are atomic
[09:52:13] <blaizco> Yes, i know, but i am talking about another subject but right
[09:53:37] <kali> as a matter of fact, you can also do it with the aggregation framework. it would be slightly more efficient than $where in terms of cpu, but the same in terms of optimisations
[10:43:33] <aandy> hi, i'm trying to query based on the existance of a field, http://pastie.org/private/gs9kg9wlapo7xhoqukpg#1-2,13,21,28,35,43,50,56,64,71,76,84 but i can't seem to figure out how to make a proper index for it, any ideas as to what i'm doing wrong?
[10:43:57] <aandy> doc snippit at the top. should i rather define the field for all docs and check for null instead?
[10:44:04] <aandy> i'm a bit of a mongodb index novice, so please bare with me ;)
[10:46:31] <aandy> (mongod 2.2.0, so indices on $exists should be ok)
[10:49:35] <aandy> querying with {$ne: null} works, http://pastie.org/private/7btc878svla8mg0uhleg#6,14 but i'm not entirely sure why $exists would scan all objects even if there's no (indexed) value for that doc
[11:09:39] <kali> the index will only works if by scanning a slice of it, you get the documents in the right order
[11:10:13] <kali> this is not the case here, as the timestamp order is disrupted by the "links" value
[11:11:23] <kali> you can try just the opposite: { timesstamp:1, links:1}. this will define an index where docs are in the right order, and i think the $exists clause will be evaluated against the index, so it should be fast