[jira] [Commented] (PHOENIX-3443) Support different encoding schemes (BYTE, SHORT, INTEGER) for storing encoded column qualifiers

["James Taylor (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/PHOENIX-3443?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15638100#comment-15638100 ] 

James Taylor commented on PHOENIX-3443:
---------------------------------------

I think the encoding scheme should be specified by a number 0-4 that indicates how many bytes will be used to encode the column qualifier. A value of 0 would indicate that the feature is disabled. Otherwise, we'd detect when the column number value goes beyond the max value based on the number of bytes (we can just use a long when we do the arithmetic), and in that case throw an error. Eventually we could potentially support a kind of clean-up procedure that would reclaim unused columns, but that could be future work if we see demand for it. The poor mans alternative would be to create a new table and do an upsert select into it.

> Support different encoding schemes (BYTE, SHORT, INTEGER) for storing encoded column qualifiers
> -----------------------------------------------------------------------------------------------
>
>                 Key: PHOENIX-3443
>                 URL: https://issues.apache.org/jira/browse/PHOENIX-3443
>             Project: Phoenix
>          Issue Type: Sub-task
>            Reporter: Samarth Jain
>




--
This message was sent by Atlassian JIRA
(v6.3.4#6332)