[
https://issues.apache.org/jira/browse/PHOENIX-3427?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15645379#comment-15645379
]
Josh Mahonin commented on PHOENIX-3427:
---------------------------------------
Thanks for the patch [~nico.pappagianis]!
Overall it looks good to me. If I can suggest an improvement, maybe also
include reading the 'TenantId' (and/or 'tid'?) from the parameters for the
DataFrame relation here:
https://github.com/apache/phoenix/blob/master/phoenix-spark/src/main/scala/org/apache/phoenix/spark/DefaultSource.scala#L39-L48
That way the syntax could support something like:
{noformat}
df.write
.format("org.apache.phoenix.spark")
.mode("overwrite")
.option("table", "TABLE1")
.option("TenantID", "theTenant")
.option("zkUrl", "localhost:2181")
.save()
{noformat}
> rdd.saveToPhoenix gives table undefined error when attempting to write to a
> tenant-specific view (TenantId defined in configuration object and passed to
> saveToPhoenix)
> -----------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: PHOENIX-3427
> URL: https://issues.apache.org/jira/browse/PHOENIX-3427
> Project: Phoenix
> Issue Type: Bug
> Reporter: Nico Pappagianis
>
> Although we can read from a tenant-specific view by passing TenantId in the
> conf object when calling sc.phoenixTableAsRDD the same does not hold for
> rdd.saveToPhoenix. Calling saveToPhoenix with a tenant-specific view as the
> table name gives a table undefined error, even when passing in the TenantId
> with the conf object.
> It appears that TenantId is lost during the execution path of saveToPhoenix.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
