[
https://issues.apache.org/jira/browse/PIG-63?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12562192#action_12562192
]
Olga Natkovich commented on PIG-63:
-----------------------------------
Ran a simple performance test:
A = load '/user/pig/tests/data/singlefile/studenttab20m' using PigStorage('\t');
store A into utf8;
The input file contains 20M records and is 419 MB. The test ran on 10 machine
cluster.
Run tests 3 times with old and new code and avergaged the results.
The results:
old code = 47.5 s
new code = 50 s
the query slowed down about 5%.
I think it is an acceptable slowdown of the correctness we get from it. Also,
the impact for more computationally intensive queries would be significantly
less.
> PigStorage does not properly handle UTF8 data
> ---------------------------------------------
>
> Key: PIG-63
> URL: https://issues.apache.org/jira/browse/PIG-63
> Project: Pig
> Issue Type: Bug
> Reporter: Olga Natkovich
> Assignee: Benjamin Reed
> Attachments: utf8.patch, utf8.patch, utf8.patch, utf8_v4.patch
>
>
> From Ben:
> I just checked the code and the problem seems to be PigStorage. getNext() uses
> readLine() which does not handle UTF8 correctly. putNext() also uses default
> encoder rather than UTF8 explicitly.
> Internally and in BinStorage UTF8 appears to be handled correctly.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
