See CASSANDRA-293 and CASSANDRA-494. Key-range updates touch multiple rows and cannot be performed atomically, so the consistency promise you get would be the same as with batch_mutate. Another other problem is that the update needs to be sent to *any* node that is responsible for part of the range. That isn't so bad though. The real problem is dealing with failure. The current implementation of hinted-handoff is key based. I've looked at extending it for ranges and there isn't a good way of doing it without rewriting all of hinted-handoff.
If you're mainly interested in avoid round-trips, have you tried using batch_mutate? I know there was a limitation when specifying a Deletion, but I think that has been resolved. On Wed, Jun 23, 2010 at 17:39, Carlos Sanchez <carlos.sanc...@riskmetrics.com> wrote: > Would there be any support for a KeyRange (start/end key) deletion in 0.7? > > Carlos > > This email message and any attachments are for the sole use of the intended > recipients and may contain proprietary and/or confidential information which > may be privileged or otherwise protected from disclosure. Any unauthorized > review, use, disclosure or distribution is prohibited. If you are not an > intended recipient, please contact the sender by reply email and destroy the > original message and any copies of the message as well as any attachments to > the original message. >
