You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@crunch.apache.org by "Chao Shi (JIRA)" <ji...@apache.org> on 2013/09/18 08:00:57 UTC

[jira] [Created] (CRUNCH-267) Fix several HFileUtils#scanHFiles related problems

Chao Shi created CRUNCH-267:
-------------------------------

             Summary: Fix several HFileUtils#scanHFiles related problems
                 Key: CRUNCH-267
                 URL: https://issues.apache.org/jira/browse/CRUNCH-267
             Project: Crunch
          Issue Type: Bug
            Reporter: Chao Shi


This patch fixes several problems about HFileUtils#scanHFiles that are discovered on our production cluster.

1. The usage of "scanner.seek" is wrong

Returning -1 indicating all KVs in the HFile is greater than the given key, so we should continue to scan. So I replaced it with seekAtOrAfter, which is copied from HBase code, and added a few tests (testScanFiles_startRow{IsTooSmall, IsTooLarge, DoesNotExist) to cover this.

2. The default implementation of HFileSource#getSize does not estimate correctly the size of input, if the input HFiles are in sub-directory (i.e. input/family/hfile)

3. There are some tricky cases about Delete/DeleteColumn. I added some test cases and fix related code. (Hopefully my test case can cover this.)

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira