[ 
https://issues.apache.org/jira/browse/CASSANDRA-3147?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

T Jake Luciani updated CASSANDRA-3147:
--------------------------------------

       Reviewer: xedin
    Description: 
The new secondary index API needs to throw IOException on some of the methods 
(where they're missing).

For index types that index an entire row (like lucene) we need to change the 
api to handle one instance of the index per ColumnFamily.  This can be done by 
adding a PerColumnSecondary and PerRowSecondaryIndex base classes.  When the 
index manager sees this it will add all columns of the same index class to the 
same index instance.  and on search it will send the entire row to the class vs 
updating each column.

  was:
The new secondary index API needs to throw IOException on some of the methods 
(where they're missing).

For index types that index an entire row (like lucene) we need to change the 
api to handle one instance of the index per ColumnFamily.  This can be done by 
adding a isRowLevelIndex() method.  When the index manager sees this it will 
add all columns of the same index class to the same index instance.  and on 
search it will send the entire row to the class vs updating each column.


> Add rowlevel support to secondary index API and methods need to throw 
> exceptions
> --------------------------------------------------------------------------------
>
>                 Key: CASSANDRA-3147
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3147
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Core
>            Reporter: Jason Rutherglen
>            Assignee: Jason Rutherglen
>            Priority: Minor
>             Fix For: 1.0.0
>
>         Attachments: 
> v1-0001-CASSANDRA-3147-add-rowLevel-logic-and-exceptions.txt
>
>
> The new secondary index API needs to throw IOException on some of the methods 
> (where they're missing).
> For index types that index an entire row (like lucene) we need to change the 
> api to handle one instance of the index per ColumnFamily.  This can be done 
> by adding a PerColumnSecondary and PerRowSecondaryIndex base classes.  When 
> the index manager sees this it will add all columns of the same index class 
> to the same index instance.  and on search it will send the entire row to the 
> class vs updating each column.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Reply via email to