[
https://issues.apache.org/jira/browse/PHOENIX-7466?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17899770#comment-17899770
]
guluo commented on PHOENIX-7466:
--------------------------------
Thank you for your response.
In our enviroment, Spark enabled authorization by Kyuubi Spark AuthZ Plugin.
And I found that the authorization service get the database information by
tableName, as follow.
https://github.com/apache/kyuubi/blob/master/extensions/spark/kyuubi-spark-authz/src/main/scala/org/apache/kyuubi/plugin/spark/authz/PrivilegesBuilder.scala#L96L100
```scala
case u if u.nodeName == "UnresolvedRelation" =>
val parts = invokeAs[String](u, "tableName").split("\\.")
val db = quote(parts.init)
val table = Table(None, Some(db), parts.last, None)
privilegeObjects += PrivilegeObject(table)
```
So, for table with default schema, we cannot the database information by
tableName in here.
> [phoenix-connector] spark with authorization enabled cannot access phoenix
> tables without schema
> ------------------------------------------------------------------------------------------------
>
> Key: PHOENIX-7466
> URL: https://issues.apache.org/jira/browse/PHOENIX-7466
> Project: Phoenix
> Issue Type: Bug
> Components: connectors
> Reporter: guluo
> Priority: Major
>
> When spark enabled authorization, Spark cannot access phoenix tables without
> schema.
> Assuming we use userA to operate Spark, and userA has access to the default
> namespace on hbase
> When Spark enabled authorization by Kyuubi Spark AuthZ Plugin, and spark
> access phoenix tables by the following way.
> df = spark.sqlContext.read.format("phoenix")
> .options(Map("table" -> "Table1", "zkUrl" -> "phoenix-server:2181")).load;
> This phoenix table doesnot specify schema, So the authorization service
> cannot get the database name of this table (database is null), So it think
> that we are accessing an unauthorized database and denies access.
> I'm not sure if it must be a problem with phoenix-connector.
> But my solution is as follows.
> When the accessed Phoenix table does not have a schema, I convert it to
> default.table in PhoenixTable.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
