[ 
https://issues.apache.org/jira/browse/ARROW-16590?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Weston Pace updated ARROW-16590:
--------------------------------
    Description: 
We've built up a number of utilities that are based around a row-major 
encoding.  Understanding these utilities requires a pretty in-depth 
understanding of this encoding format.

In addition, as we start to have external work done on exec nodes I think some 
of these utilities will be useful to eventually expose externally.

This JIRA is for work to refactor the existing row-based utilities and 
consolidate them into a separate location.

  was:
We've built up a number of utilities that are based around a row-major 
encoding.  Understanding these utilities requires a pretty in-depth 
understanding of this encoding format.

In addition, as we start to have external work done on exec nodes I think some 
of these utilities will be useful to eventually expose externally.

This JIRA attempts to address both goals purely through renaming & moving.  It 
does not entirely accomplish the second as that would require converting some 
types to PIMPL.  However, it does establish a pretty solid groundwork should we 
decide to later do that.

In addition, this adds tests to the external surface of these APIs, something 
that had only previously been tested indirectly via the hash join.  Eventually 
I suspect we will also want benchmarks around these utilities but I will save 
that for a future PR.


> [C++] Consolidate files dealing with row-major storage, add some helper 
> methods
> -------------------------------------------------------------------------------
>
>                 Key: ARROW-16590
>                 URL: https://issues.apache.org/jira/browse/ARROW-16590
>             Project: Apache Arrow
>          Issue Type: Sub-task
>          Components: C++
>            Reporter: Weston Pace
>            Priority: Major
>              Labels: pull-request-available
>          Time Spent: 1h 20m
>  Remaining Estimate: 0h
>
> We've built up a number of utilities that are based around a row-major 
> encoding.  Understanding these utilities requires a pretty in-depth 
> understanding of this encoding format.
> In addition, as we start to have external work done on exec nodes I think 
> some of these utilities will be useful to eventually expose externally.
> This JIRA is for work to refactor the existing row-based utilities and 
> consolidate them into a separate location.



--
This message was sent by Atlassian Jira
(v8.20.7#820007)

Reply via email to