I think Jens provided the link to the replication "How". I might then just first tag your bad documents in the old DB - then filter replicate to prevent having to build some giant map in a filter function. I've followed this approach in the past and found it effective.
- Jim Sent from my iPhone On May 21, 2012, at 8:12 AM, "Tim Tisdall" <tisd...@gmail.com> wrote: > I'm not sure exactly what is or how to filter replicate, someone have a > link to a page describing how to do this? I have a feeling, though, that > I wouldn't be able to do it because the list of the specific ids I needed > to remove were in a sqlite database and there was no way to determine the > "bad documents" solely on their content. > > -Tim > > On Mon, May 21, 2012 at 12:03 AM, Paul Davis > <paul.joseph.da...@gmail.com>wrote: > >> Yeah, likely the best way to undo this would be to use filtered >> replication to a local db and then rename the .couch files and reboot >> to get it swapped over. >> >> On Fri, May 18, 2012 at 1:02 PM, Jim Klo <jim....@sri.com> wrote: >>> Not sure how many docs you have, but can you filter replicate the good >> docs >>> into a new db? >>> >>> >>> Jim Klo >>> Senior Software Engineer >>> Center for Software Engineering >>> SRI International >>> >>> On May 18, 2012, at 10:39 AM, Tim Tisdall wrote: >>> >>> After further reading it seems like I could use _purge ... However, I >>> still need to query the DB to fetch all the revisions. >>> >>> On Fri, May 18, 2012 at 12:12 PM, Tim Tisdall <tisd...@gmail.com> wrote: >>> >>> I accidentally created several million documents in my DB which I'd like >>> >>> to undo. ^_^ I know I can query each document to get the revision and >>> >>> then issue a DELETE, but that seems like a lot of additional work. Also, >>> >>> the wiki says "Deleted documents remain in the database forever, even >> after >>> >>> compaction, to allow eventual consistency when replicating." and that's a >>> >>> lot of dead space. Is there a way that I can purge all of those >> documents >>> >>> given that I have the _ids that they were saved under? I essentially >> would >>> >>> like to return the DB to a state before I inserted all of those >> documents. >>> >>> >>> I'm also going to be inserting updated documents to the DB with those >> same >>> >>> _ids, so another alternative is to post updates to each (but that would >>> >>> also require getting the revision ids) and then clearing out old >> revisions. >>> >>> However, this would require fetching several million revision ids and >> then >>> >>> figuring out how to force the DB to clear out all old revisions. >>> >>> >>> -Tim >>> >>> >>> >>