[
https://issues.apache.org/jira/browse/CASSANDRA-626?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12789168#action_12789168
]
Paul R. Brown commented on CASSANDRA-626:
-----------------------------------------
I think the bug is invalid but still an interesting comment: I misinterpreted
the documentation about how columns are sorted and confused that with how rows
are sorted. Sorting by Long type only applies to columns, so the bug is not
well-posed.
I'm going to change it to documentation bug. Some coverage of sorting for key
names would be worthwhile, but the reader can make the assumption that keys are
sorted in Java sort order for Strings.
> Polymorphic status of keys as byte[] and String is confusing
> ------------------------------------------------------------
>
> Key: CASSANDRA-626
> URL: https://issues.apache.org/jira/browse/CASSANDRA-626
> Project: Cassandra
> Issue Type: Bug
> Components: Core, Documentation & website
> Affects Versions: 0.5
> Reporter: Paul R. Brown
>
> As part of some experiments with Cassandra, I wanted to build a column space
> with row keys that are Long (64-bit binary blocks) and column names that are
> Long, and this seems to be supported by the data configuration infrastructure:
> {code}<ColumnFamily ColumnType="Super"
> CompareWith="LongType"
> CompareSubcolumnsWith="LongType"
> Name="UserEvent"/>{code}
> But there is no obvious way to perform the insert over the Thrift interface
> via Java; I tried various ways of turning the Long bits into Strings for the
> API call, but the insert call returns:
> {code}Exception in thread "main" InvalidRequestException(why:A long is
> exactly 8 bytes){code}
> The status of keys as either byte[] or String and the mechanisms for
> inserting need to be clarified and surfaced in the API.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
