Re: Map/Reduce data order

2010-04-15 Thread Jimmy Thrasher
Hey Sean, After reading and trying a bit more, I was reminded of the idempotency requirement for reduce functions, so even if the results ended up in order, I wouldn't be "allowed" to count on it. Thanks for the clarification, Jimmy On Thu, Apr 15, 2010 at 9:13 AM, Sean Cribbs wrote: > Jimmy,

Re: Map/Reduce data order

2010-04-15 Thread Sean Cribbs
Jimmy, Consider map phases to be completely parallel and unordered. When a map phase completes, the results are sent back to the coordinating node, which -- assuming the next phase is also a map phase -- will send them out to the vnodes that own the new set, which will perform their map phase

Map/Reduce data order

2010-04-14 Thread Jimmy Thrasher
Hi folks, I have a query where one of the map steps spits out the original entry (via [v.bucket, v.key]) and a derivative bucket,key pair. If the original key is "asdf" the result of the map step would look like this: [["events", "asdf"], ["asdf,annotation", "some-annotation-key"]] Future map st