[
https://issues.apache.org/jira/browse/CASSANDRA-3089?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13092324#comment-13092324
]
Rick Shaw commented on CASSANDRA-3089:
--------------------------------------
The JDBC {{RowId}} assumes it can be compared with another {{RowId}} for
identity. The stated specification says that they are the same if they have the
same logical/physical identity and they come from the same Table and Dataset.
This implies the same CF and KS as well as the matching row index value. To
illustrate: if a rows in two CFs "A", and "B" had {{Integer}} indexes, it could
easily happen that both CFs had a row with the index value of integer 1. But a
{{RowId}} from CF "A" containing 1 and a {{RowId}} containing 1 from CF "B"
should not report being equal because they come from different CFs.
However it is not common practice to store the CF and the KS along with a row
index when creating secondary indexes either ad-hoc or through C* internal
tooling. The KS can be implied as the current KS because it is not useful in C*
for it to be anything but the current KS, although I guess it is _possible_ as
well. But the CF can not really be known unless we create a new {{RowIdType}},
and that seems like it would never really be used in common practice anyway.
The alternative is to document that we do not really comply, and just use
equality of the byte value contents of the contained in the column's value.
> Support RowId in ResultSet
> --------------------------
>
> Key: CASSANDRA-3089
> URL: https://issues.apache.org/jira/browse/CASSANDRA-3089
> Project: Cassandra
> Issue Type: Sub-task
> Components: Drivers
> Affects Versions: 0.8.4
> Reporter: Rick Shaw
> Assignee: Rick Shaw
> Priority: Trivial
> Labels: JDBC, lhf
> Fix For: 1.0
>
>
> Support the JDBC concept of {{RowId}} by using the C* row index value.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
