You are viewing a plain text version of this content. The canonical link for it is here.
Posted to notifications@accumulo.apache.org by "Christopher Tubbs (JIRA)" <ji...@apache.org> on 2019/04/22 20:02:00 UTC

[jira] [Resolved] (ACCUMULO-394) Provide framework for combining multiple columns

     [ https://issues.apache.org/jira/browse/ACCUMULO-394?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Christopher Tubbs resolved ACCUMULO-394.
----------------------------------------
    Resolution: Won't Fix

This can already be done with iterators more generally, rather than special-purpose combiners. If this is still an issue, please open an issue on GitHub, where we now track issues: https://github.com/apache/accumulo/issues

> Provide framework for combining multiple columns
> ------------------------------------------------
>
>                 Key: ACCUMULO-394
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-394
>             Project: Accumulo
>          Issue Type: New Feature
>            Reporter: John Vines
>            Priority: Trivial
>
> Currently the combiner frameworks provides mechanisms for users to reduce the values of all Keys down to the column visibility label. I imagine there are cases where users want to combine all values for all visibilities of a column family, or even a row. Perhaps even just combine visibilities. Also, to extend combination to those Key elements instead of just values. There's a lot of ways to generalize this problem space, and I think it's something worth looking into. I think we should be able to use something similar to the current Combiner framework, except the reduce function's input Key is a partial key for the amount of uniqueness being saught. The majority of it should be straight forward, with the exception of visibilities which should be logically ANDed together (but this may be something we want to defer to the combiner's writer, I don't know).



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)