[
https://issues.apache.org/jira/browse/CASSANDRA-11354?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15204301#comment-15204301
]
Benjamin Lerer commented on CASSANDRA-11354:
--------------------------------------------
{quote}Is AbstractSingleRestriction still required as an abstract class? It's
also possible to implement the false returning methods on the interface
directly. {quote}
I removed the {{AbstractSingleRestriction}} class and moved the
{{reverseBoundIfNeeded}} method into the {{Bound}} class.
{quote}There are two places where clustering column restrictions are
"validated"{quote}
The main reason for that is that the restrictions are needed in order for the
second validation. When merging restrictions we have no guaranty of the order
in which the restrictions are provided. It could be {{SELECT * FROM myTable
WHERE clustering1 = 1 AND clustering2 = 2;}} or {{SELECT * FROM myTable WHERE
clustering2 = 2 AND clustering1 = 1;}}
> PrimaryKeyRestrictionSet should be refactored
> ---------------------------------------------
>
> Key: CASSANDRA-11354
> URL: https://issues.apache.org/jira/browse/CASSANDRA-11354
> Project: Cassandra
> Issue Type: Improvement
> Components: CQL
> Reporter: Benjamin Lerer
> Assignee: Benjamin Lerer
>
> While reviewing CASSANDRA-11310 I realized that the code of
> {{PrimaryKeyRestrictionSet}} was really confusing.
> The main 2 issues are:
> * the fact that it is used for partition keys and clustering columns
> restrictions whereas those types of column required different processing
> * the {{isEQ}}, {{isSlice}}, {{isIN}} and {{isContains}} methods should not
> be there as the set of restrictions might not match any of those categories
> when secondary indexes are used.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
