You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@orc.apache.org by "Owen O'Malley (Jira)" <ji...@apache.org> on 2021/02/17 00:15:00 UTC

[jira] [Commented] (ORC-618) Make several methods in TreeReader public

    [ https://issues.apache.org/jira/browse/ORC-618?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17285557#comment-17285557 ] 

Owen O'Malley commented on ORC-618:
-----------------------------------

I'd recommend using the org.apache.orc.util.StreamWrapperFileSystem. It was created exactly for this purpose.

You might also be interested in ORC-508, which introduces a new API that does not depend on Hadoop classes.

> Make several methods in TreeReader public
> -----------------------------------------
>
>                 Key: ORC-618
>                 URL: https://issues.apache.org/jira/browse/ORC-618
>             Project: ORC
>          Issue Type: Improvement
>            Reporter: Pasha Finkeshteyn
>            Priority: Major
>
> I've implemented my own RecordReader and DataReader according to ORC-617 — based on only FSDataInputStream, not Path. Of course, I've implemented them being heavily influenced by RecordReaderImpl and DefaultDataReader (yes, it was almost copy-paste which only says us that it should be absolutely possible to implement them with only FSDataInputStream).
>  
> But during implementation I've found that methds skipRows, setVectorColumnCount and startStripe of TreeReader are package-private. I can't see a reason for them not to be public and now I have to call them with reflection which looks weird and creepy.
>  
> Please, change visibility of this methods. It shouldn't break any existing functionality for current adopters and will improve experience for new developers.
>  
> Thanks!



--
This message was sent by Atlassian Jira
(v8.3.4#803005)