sqoop-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Daniel Voros (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (SQOOP-3267) Incremental import to HBase deletes only last version of column
Date Tue, 23 Jan 2018 16:12:00 GMT

    [ https://issues.apache.org/jira/browse/SQOOP-3267?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16335982#comment-16335982
] 

Daniel Voros commented on SQOOP-3267:
-------------------------------------

[~vasas], thanks for your reply!

The current implementation does in fact issue separate Put commands, for every column, but
I've already addressed this issue in  [^SQOOP-3267.1.patch] (see first comment on this issue).
There is still a slight overhead in adding all columns to the Put command but it's way better
then 1 Put/column.

Despite this I still agree with you on providing the option to configure the behavior (ignore/delete/null-string).
However, knowing it's only a small performance overhead, we could as well make null-string
the default.

All right, let's introduce --hbase-null-string and do not overload --null-string.

If you agree with making it the default, I think I'd implement the null-string mode in this
ticket and open a new Jira for the configurability of ignore/delete/null-string and --hbase-null-string.

> Incremental import to HBase deletes only last version of column
> ---------------------------------------------------------------
>
>                 Key: SQOOP-3267
>                 URL: https://issues.apache.org/jira/browse/SQOOP-3267
>             Project: Sqoop
>          Issue Type: Bug
>          Components: hbase-integration
>    Affects Versions: 1.4.7
>            Reporter: Daniel Voros
>            Assignee: Daniel Voros
>            Priority: Major
>         Attachments: SQOOP-3267.1.patch
>
>
> Deletes are supported since SQOOP-3149, but we're only deleting the last version of a
column when the corresponding cell was set to NULL in the source table.
> This can lead to unexpected and misleading results if the row has been transferred multiple
times, which can easily happen if it's being modified on the source side.
> Also SQOOP-3149 is using a new Put command for every column instead of a single Put per
row as before. This could probably lead to a performance drop for wide tables (for which HBase
is otherwise usually recommended).
> [~jilani], [~anna.szonyi] could you please comment on what you think would be the expected
behavior here?



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Mime
View raw message