[
https://issues.apache.org/jira/browse/CASSANDRA-344?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jonathan Ellis updated CASSANDRA-344:
-------------------------------------
Attachment: 344-v4.patch
Made a couple changes:
- renamed thrift method to get_range_slice for consistency [my fault, sorry]
- renamed some things to RangeSliceX instead of SliceX or SliceRangeX to
satisfy my OCD
- RangeSliceReply now uses Rows internally [my fault for not mentioning that
too], and propagates rangeCompletedLocally
- added CFS.getKeyRangeRaw to do a keyrange w/o filtering out deleted rows.
avoiding that behavior is one of the reasons we're adding RangeSlice.
Still todo:
- use rangeCompletedLocally in StorageProxy.getKeySlice (oops, missed that one
-- rename to getRangeSlice :)
- handle both kinds of column slice predicates, right now it assumes it is
always a range. i suggest leaving SlicePredicate un-flattened in
RangeSliceCommand. (in fact the way it is now it is guaranteed to NPE since
exactly one of the Predicate sub-structs should be null.) Look at
SliceByNamesReadCommand and SliceFromReadCommand for turning the predicate into
a QueryFilter.
- add a test to test/system/test_server.py exercising both kinds of predicates
to get_range_slice
Optionally, you might want to do this to serialize the Predicate directly:
http://wiki.apache.org/thrift/ThriftUsageJava . Of course there is an extra
copy involved. (We don't actually use that in Cassandra anywhere yet. We
probably should.)
Almost done! :)
> allow slicing during key range query
> ------------------------------------
>
> Key: CASSANDRA-344
> URL: https://issues.apache.org/jira/browse/CASSANDRA-344
> Project: Cassandra
> Issue Type: New Feature
> Components: Core
> Reporter: Jonathan Ellis
> Assignee: Gary Dusbabek
> Fix For: 0.5
>
> Attachments: 344-thrift.patch, 344-v1.patch, 344-v2.patch,
> 344-v3.diff, 344-v4.patch
>
>
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
