[
https://issues.apache.org/jira/browse/SQOOP-1744?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14232297#comment-14232297
]
Ryan Blue commented on SQOOP-1744:
----------------------------------
You're right that HBase doesn't buy us much. In that situation, where we can't
isolate a subset of the data that might change, I think we have two options:
either rewrite the entire dataset each time or maintain the dataset in HBase.
We shouldn't overlook the second option, which would facilitate the fetch
frequency that you want. Parquet is a great format to use, but if we have to
constantly rewrite the entire dataset or very substantial portions of it, then
it might not be worth the storage savings.
> TO-side: Write data to HBase
> ----------------------------
>
> Key: SQOOP-1744
> URL: https://issues.apache.org/jira/browse/SQOOP-1744
> Project: Sqoop
> Issue Type: Sub-task
> Components: connectors
> Reporter: Qian Xu
> Assignee: Qian Xu
> Fix For: 1.99.5
>
>
> Propose to write data into HBase. Note that different to HDFS, HBase is
> append only. Merge does not work for HBase.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
