On Fri, Mar 27, 2015 at 9:03 AM, Luca Morandini <lmorand...@ieee.org> wrote: > > Hope this helps. >
Thinking about every example at this point is helpful, eve if it doesn't solve my immediate problem. Thanks everyone for the replies. I'm currently thinking of combining two views together to solve different parts of the problem: One that will give me a list of Item IDs that have changed recently, and then I can pass that array to the other view with a keys= array. Not ideal, but seems to work pretty fast. Note: I can better document my specific data/etc if people are curious, but trying to avoid flooding the list with that level of detail. Possibly the closest analogy would be to say that I have "items" which have increasing version numbers and I have "stores", and any given item can be in any store. I have an Item_store document type which has details about a specific item and its version in a store. I have a _view which has a map that emits (item, [item.revision, store]) , and a reduce that finds the largest revision, and then returns an array of all the stores that have the largest revision. I then have a _list which iterates through all the keys (one per unique item) and skips the ones for a store I gave it (IE: I'm trying to find all the items which are in some other store, but not this one). In most cases every store already has the latest revision of every item, so I'm looking for a rare scenario. I tried having a per-store view, but since every key would still exist (and most would just be NULL meaning that the store already had the latest item), there was no saved time over having a generic view that would then have a _list that would filter. -- System Administration and software developer, Canadiana.org http://www.canadiana.ca
