cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Nate McCall (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-5970) FilteredRangeSlice command for regex searches against column names on known sets of keys
Date Tue, 03 Sep 2013 14:48:52 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-5970?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13756668#comment-13756668
] 

Nate McCall commented on CASSANDRA-5970:
----------------------------------------

We do, but I don't think anything in our current implementation precludes adding functionality
to the existing slice plumbing for the actual regex column matching. And I agree it would
be cleaner to have it tucked in to what is there. 

That harder part - perhaps just to me as I'm still coming back up to speed on internals -
will be where/how to apply the key filtering. Would this be adding a 'KeyFilter' member (as
described above) and another overloaded c-tor for such to RangeSliceCommand (similar to how
Bounds is passed in get_[range|paged]_slices delegation to StorageProxy)?
                
> FilteredRangeSlice command for regex searches against column names on known sets of keys
> ----------------------------------------------------------------------------------------
>
>                 Key: CASSANDRA-5970
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-5970
>             Project: Cassandra
>          Issue Type: New Feature
>          Components: Core
>            Reporter: Nate McCall
>
> This is the ability to apply a regex against columns when the set of keys is known. In
filtering the keys, we would like to allow for the following clauses: E, GTE, LTE, NE, inclusive
list, exclusive list.
> The end goal is to provide for efficient searching in the case where you have some knowledge
of the keys. A specific use case would be, say, searching user agent strings in the given
set of date buckets in the classic time-series web log use case. This is a "sweet spot" for
Cassandra and providing a more direct method of access for such will help a lot of users.
> Additionally, this will provide some level of feature parity with RDBMS crowd who've
had this feature for some time.
> Internally, this will include the introduction of a new Verb, SSTableScanner extension
and an ExtendedFilter implementation which applies the regex as well as a new method on StorageProxy.
> This issue does not cover exposing this new query method to thrift and CQL, but obviously
that will be required for this to be of any practical use. Those should be covered by separate
issues.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message