[
https://issues.apache.org/jira/browse/HIVE-5728?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13883892#comment-13883892
]
Lefty Leverenz commented on HIVE-5728:
--------------------------------------
Review of comments in HiveConf.java & descriptions in hive-default.xml.template
--
hive.exec.orc.default.stripe.size
* Comment & description should specify units (bytes): "Define the default ORC
stripe size."
hive.exec.orc.default.row.index.stride
* Comment & description say "stripe" instead of "stride": "Define the default
ORC index stripe."
* Should explain that stride is the number of rows between index entries.
(Stripes contain as many strides as fit in that size, if I understand the
wikidoc correctly.)
* Default value is different in comment (null) and description (10000).
hive.exec.orc.default.buffer.size
* Default value is different in comment (null) and description (262144).
* Should specify units (presumably bytes).
hive.exec.orc.default.block.padding
* Default value is different in comment (null) and description (true).
* Would be good to explain block padding, either here or in the wiki: "Define
the default block padding."
hive.exec.orc.default.compress
* Comment needs all-caps ORC: "Define the default orc compress" (nitpickers R
us) but better to use the definition's wording: "Define the default
compression codec for ORC file."
* Default value is different in comment (null) and description (ZLIB).
hive.exec.orc.dictionary.key.size.threshold
* Looks like you wanted to delete its one-line entry in HiveConf.java, then add
it below the other configs on two lines -- but instead you've deleted a blank
line so now it's in there twice.
* How about adding a comment (copying the definition in
hive-default.xml.template)?
> Make ORC InputFormat/OutputFormat usable outside Hive
> -----------------------------------------------------
>
> Key: HIVE-5728
> URL: https://issues.apache.org/jira/browse/HIVE-5728
> Project: Hive
> Issue Type: Improvement
> Components: File Formats
> Reporter: Daniel Dai
> Assignee: Daniel Dai
> Fix For: 0.13.0
>
> Attachments: HIVE-5728-1.patch, HIVE-5728-10.patch,
> HIVE-5728-2.patch, HIVE-5728-3.patch, HIVE-5728-4.patch, HIVE-5728-5.patch,
> HIVE-5728-6.patch, HIVE-5728-7.patch, HIVE-5728-8.patch, HIVE-5728-9.patch,
> HIVE-5728.10.patch
>
>
> ORC InputFormat/OutputFormat is currently not usable outside Hive. There are
> several issues need to solve:
> 1. Several class is not public, eg: OrcStruct
> 2. There is no InputFormat/OutputFormat for new api (Some tools such as Pig
> need new api)
> 3. Has no way to push WriteOption to OutputFormat outside Hive
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
