[
https://issues.apache.org/jira/browse/PHOENIX-2429?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15011562#comment-15011562
]
Josh Mahonin commented on PHOENIX-2429:
---------------------------------------
You're right, the properties are getting copied over from the Configuration
object to the Properties object for both the input and output connections. And
for the Spark integration, the Configuration object is probably the only good
way to pass in custom settings like these as well.
It looks like the issue is a naming mismatch. There's
PhoenixConfigurationUtil.CURRENT_SCN_VALUE ("phoenix.mr.currentscn.value") and
PhoenixRuntime.CURRENT_SCN_ATTRIB ("CurrentSCN"), and that code you have quoted
above does the translation between the two on the PhoenixInputFormat only. On
the output side, that translation step is missing.
I think there's a workaround, which is to set 'CurrentSCN' on the configuration
object, instead of 'phoenix.mr.currentscn.value', but I think it would be a
good idea to standardize on one config setting.
> PhoenixConfigurationUtil.CURRENT_SCN_VALUE for phoenix-spark plugin does not
> work
> ---------------------------------------------------------------------------------
>
> Key: PHOENIX-2429
> URL: https://issues.apache.org/jira/browse/PHOENIX-2429
> Project: Phoenix
> Issue Type: Bug
> Affects Versions: 4.2.0, 4.6.0
> Reporter: Diego Fustes Villadóniga
>
> When I call the method saveToPhoenix to store the contents of a ProductDD,
> passing a hadoop configuration, where I set
> PhoenixConfigurationUtil.CURRENT_SCN_VALUE to be a given timestamp, the
> values are not stored with such timestamp, but using the server time
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
