[
https://issues.apache.org/jira/browse/HADOOP-2339?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12550676
]
Bryan Duxbury commented on HADOOP-2339:
---------------------------------------
Let's specify very clearly what this issue regards.
As I understand it, the goal is to be able to delete a specific column family
and qualifier on every row in the entire table. It's important to note that the
qualifier is needed, because you can already do this essentially with ALTER
TABLE if you want to delete an entire column family. Is this correct?
If so, I'm not sure that we should offer this functionality. It will be slow,
especially if it's run as a single delete at a time in serial from the shell. I
would think that from the shell, you have the ability to empty an entire table
with TRUNCATE, or delete an entire column family with ALTER TABLE ... DROP
columnfamily, and that's probably enough. If you actually do need to delete
just a given qualified column name on every row in your table, you should write
a mapreduce job or some other highly parallel way to do that. This is in
keeping with the goal of making the shell both efficient and lean when it comes
to the functionality it supports.
> [Hbase Shell] Delete command with no WHERE clause
> -------------------------------------------------
>
> Key: HADOOP-2339
> URL: https://issues.apache.org/jira/browse/HADOOP-2339
> Project: Hadoop
> Issue Type: Improvement
> Components: contrib/hbase
> Affects Versions: 0.16.0
> Reporter: Edward Yoon
> Assignee: Edward Yoon
> Fix For: 0.16.0
>
> Attachments: 2339.patch, 2339_v02.patch, 2339_v03.patch,
> 2339_v04.patch
>
>
> using HbaseAdmin.deleteColumn() method.
> {code}
> DELETE column_name FROM table_name;
> {code}
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
