[SPARK-8572] Type coercion for ScalaUDFs

• Type: Bug
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Fix Version/s: 1.5.0
• Component/s: SQL
• Target Version/s: 1.4.2, 1.5.0
• Assignee: Cheolsoo Park
• Reporter: Yin Huai
• Created: 23/Jun/15 20:13
• Updated: 04/Jul/15 05:14
• Resolved: 04/Jul/15 05:14

Description

Seems we do not do type coercion for ScalaUDFs. The following code will hit a runtime exception.

import org.apache.spark.sql.functions._
val myUDF = udf((x: Int) => x + 1)
val df = sqlContext.range(1, 10).toDF("i").select(myUDF($"i"))
df.explain(true)
df.show

It is also good to check if we do type coercion for PythonUDFs.

Issue Links

• links to: [Github] Pull Request #7203 (piaozhexiu)