Orc: Closes #5777 - Obtain ORC stripe offsets from writer

Open pavibhai opened this issue 3 years ago • 1 comments

What?

Fixes ORCFileAppender such that it can retrieve the stripe offsets from the writer instead of having to reopen the file.

From ORC 1.7 we have a public API for getting stripe information from the writer. This avoids an additional IO to get stripe offsets

Added unit test to verify that the offsets are matching the offsets read from the file

Sep 16 '22 18:09 pavibhai

@kbendick Hope you are still looking at ORC support on Iceberg. Can you please take a look at this PR?

Sep 22 '22 23:09 pavibhai

Looks good other than an error message and some tests for unrelated cases.

@rdblue I have addressed the review comments can you please take another look at this?

Sep 27 '22 16:09 pavibhai

Thanks, @pavibhai! Looks great so I merged it.

Sep 28 '22 20:09 rdblue