[ 
https://issues.apache.org/jira/browse/ARROW-1008?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16006834#comment-16006834
 ] 

Wes McKinney commented on ARROW-1008:
-------------------------------------

As part of this, we should make {{StreamReader}} an abstract base class with a 
{{virtual GetNextRecordBatch}} API. We may want to have other implementations 
of {{StreamReader}} to provide a uniform C++ interface to streaming Arrow data. 
The current StreamReader presumes a particular byte stream representation

> [C++] Define abstract interface for stream iteration
> ----------------------------------------------------
>
>                 Key: ARROW-1008
>                 URL: https://issues.apache.org/jira/browse/ARROW-1008
>             Project: Apache Arrow
>          Issue Type: New Feature
>          Components: C++
>            Reporter: Wes McKinney
>
> The purpose of this JIRA is to decouple the physical structure of the stream 
> from the StreamReader API. So if we wanted to put the stream components into 
> different physical files on disk, this would permit the construction of a 
> different kind of StreamIterator that knows how to read the respective stream 
> components from files



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

Reply via email to