[
https://issues.apache.org/jira/browse/ORC-101?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15507465#comment-15507465
]
ASF GitHub Bot commented on ORC-101:
------------------------------------
Github user omalley commented on the issue:
https://github.com/apache/orc/pull/60
Ok, the latest push has a few changes:
* bloom_filter_utf8 streams use a new encoding with bytes instead of
long[]. This is much more efficient for performance and storage size.
* all column types now have bloom_filter_utf8 streams (largely to get the
new representation)
* the default is to just write the new bloom_filter_utf8 streams that old
readers will ignore. There is an option to write both bloom_filter and
bloom_filter_utf8 streams to support old readers.
* there is an option for new readers to ignore the old bloom filters.
* files generated after hive-12055 will correctly use the utf8 encoding
even for the bloom_filter stream.
> Correct the use of the default charset in the bloomfilter
> ---------------------------------------------------------
>
> Key: ORC-101
> URL: https://issues.apache.org/jira/browse/ORC-101
> Project: Orc
> Issue Type: Improvement
> Reporter: Owen O'Malley
> Assignee: Owen O'Malley
>
> Currently ORC's bloom filter depends on the default character set, which
> isn't constant between computers.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
