[
https://issues.apache.org/jira/browse/HIVE-9274?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14270096#comment-14270096
]
Thomas Friedrich commented on HIVE-9274:
----------------------------------------
Looks like this behavior has changed due to HIVE-6584. In DDLTask the input and
output format classes are only persisted when explicitly specified. I wonder if
the documentation in
https://cwiki.apache.org/confluence/display/Hive/StorageHandlers should be
updated (under open issues: "Names of helper classes such as input format and
output format are saved into the metastore based on what the storage handler
returns during CREATE TABLE; it would be better to leave these null in case
they are changed later as part of a handler upgrade").
> STORED BY not setting input and output formats correctly
> --------------------------------------------------------
>
> Key: HIVE-9274
> URL: https://issues.apache.org/jira/browse/HIVE-9274
> Project: Hive
> Issue Type: Bug
> Affects Versions: 0.14.0
> Reporter: Thomas Friedrich
> Priority: Minor
>
> When creating a new table using a storage handler via the STORED BY clause,
> for example using the HBaseStorageHandler, the input and output formats are
> set to null instead of the correct formats:
> CREATE TABLE hbase_table_1(key int, value string)
> STORED BY 'org.apache.hadoop.hive.hbase.HBaseStorageHandler'
> WITH SERDEPROPERTIES ("hbase.columns.mapping" = ":key,cf1:val")
> TBLPROPERTIES ("hbase.table.name" = "xyz");
> describe formatted hbase_table_1:
> # Storage Information
> SerDe Library: org.apache.hadoop.hive.hbase.HBaseSerDe
> InputFormat: null
> OutputFormat: null
> In older Hive versions, it would set the correct formats.
> Could it be related to the changes in HIVE-5976?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
