[
https://issues.apache.org/jira/browse/KUDU-3440?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17723007#comment-17723007
]
Ranga Reddy commented on KUDU-3440:
-----------------------------------
Hi [~mgreber]
I have tested Kudu with Spark 3.3 in CDP 7.1.8 cluster and i am able to insert
the data.
{code:java}
val kudu_table = "default.employees"
val kudu_master = "kudu.master1:7051,kudu.master2:7051,kudu.master2:7051"
// Create sample dataset to insert
case class Employee(id:Long, name: String, age: Short, salary: Float)
val employeeDF = Seq(
Employee(4L, "Employee6", 56, 1500.5f),
Employee(5L, "Employee7", 30, 15000.5f)
).toDF()
// Save the data to kudu
employeeDF.write.options(Map("kudu.master"-> kudu_master, "kudu.table"->
kudu_table)).mode("append").format("kudu").save() {code}
> Add support for Spark 3.3
> -------------------------
>
> Key: KUDU-3440
> URL: https://issues.apache.org/jira/browse/KUDU-3440
> Project: Kudu
> Issue Type: Improvement
> Reporter: Marton Greber
> Priority: Minor
> Attachments: spark_3_3.patch
>
>
> As it turns out Kudu-Spark bindings are not compatible upwards of 3.2.
> The reason being is that in Spark 3.3, compareBinary() has been moved from
> org.apache.spark.sql.catalyst.util.TypeUtils to
> org.apache.spark.unsafe.types.ByteArray.
> As of now there is no support for Spark 3.3 with Kudu, because of this
> breaking change.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
