[
https://issues.apache.org/jira/browse/PHOENIX-2116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14710069#comment-14710069
]
Josh Mahonin commented on PHOENIX-2116:
---------------------------------------
Have updated the patch. Only notable changes are in the unit tests from
previous version:
Comments:
a) Unfortunately, the act of calling 'configure()' on PhoenixSink seems to
require a running cluster.
b) I've created a test which calls a mock instance of the PhoenixSink, to
verify that configure() is invoked. However, the serializer is a bit trickier,
since it's instantiated using a Class.forName(). Instead, I've made a
CustomSerializer which writes some data to Phoenix, then the test verifies the
data was written.
> phoenix-flume: Sink/Serializer should be extendable
> ---------------------------------------------------
>
> Key: PHOENIX-2116
> URL: https://issues.apache.org/jira/browse/PHOENIX-2116
> Project: Phoenix
> Issue Type: Improvement
> Affects Versions: 4.5.0, 4.4.1
> Reporter: Josh Mahonin
> Assignee: Josh Mahonin
> Attachments: PHOENIX-2116-v2.patch, PHOENIX-2116.patch
>
>
> When using flume, often times custom serializers are necessary to transform
> data before sending to a sink. The existing Phoenix implementation however
> makes it difficult to extend and add new functionality.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)