You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@arrow.apache.org by "Louis Calot (Jira)" <ji...@apache.org> on 2022/09/23 09:03:00 UTC

[jira] [Created] (ARROW-17825) [C++] Allow to write several tables successively with ORCFileWriter::Write method

Louis Calot created ARROW-17825:
-----------------------------------

             Summary: [C++] Allow to write several tables successively with ORCFileWriter::Write method
                 Key: ARROW-17825
                 URL: https://issues.apache.org/jira/browse/ARROW-17825
             Project: Apache Arrow
          Issue Type: Improvement
          Components: C++
            Reporter: Louis Calot
            Assignee: Louis Calot
             Fix For: 10.0.0


I had the need to write an ORC file little by little, so as to not consume too much memory.
Following [this|https://github.com/apache/arrow/issues/14211] discussion, it appeared that the API did not seemed to prevent doing that, but that the internal implementation was not reusing the writer accordingly.

This PR makes the needed changes to reuse the "writer_" correctly.

I do not think that the preceding behaviour was correct, as calling several time the "Write" method would lead to incorrect ORC files.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)