[
https://issues.apache.org/jira/browse/HIVE-26105?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
mahesh kumar behera resolved HIVE-26105.
----------------------------------------
Resolution: Fixed
> Show columns shows extra values if column comments contains specific Chinese
> character
> ---------------------------------------------------------------------------------------
>
> Key: HIVE-26105
> URL: https://issues.apache.org/jira/browse/HIVE-26105
> Project: Hive
> Issue Type: Bug
> Components: Hive, HiveServer2
> Reporter: mahesh kumar behera
> Assignee: mahesh kumar behera
> Priority: Major
> Labels: pull-request-available
> Time Spent: 0.5h
> Remaining Estimate: 0h
>
> The issue is happening because the UTF code for one of the Chinese character
> contains the binary value of '\r' (CR). Because of this, the Hadoop line
> reader (used by fetch task in Hive) is assuming the value after that
> character as new value and this extra value with junk is getting displayed.
> The issue is with 0x540D 名 ... The last value is "D" ..that is 13. While
> reading the result, Hadoop line reader interpreting it as CR ( '\r'). Thus an
> extra value with Junk is coming as output. For show column, we do not need
> the comments. So while writing to the file, only column names should be
> included.
> [https://github.com/apache/hadoop/blob/0fbd96a2449ec49f840d93e1c7d290c5218ef4ea/hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/util/LineReader.java#L238]
>
> {code:java}
> create table tbl_test (fld0 string COMMENT '期 ' , fld string COMMENT
> '期末日期', fld1 string COMMENT '班次名称', fld2 string COMMENT '排班人数');
> show columns from tbl_test;
> +--------+
> | field |
> +--------+
> | fld |
> | fld0 |
> | fld1 |
> | � |
> | fld2 |
> +--------+
> 5 rows selected (171.809 seconds)
> {code}
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
