[SPARK-38481] Substitute Java overflow exception from TIMESTAMPADD by Spark exception

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.3.0
• Fix Version/s: 3.3.0
• Component/s: SQL
• Assignee: Max Gekk
• Reporter: Max Gekk
• Created: 09/Mar/22 11:01
• Updated: 10/Mar/22 12:24
• Resolved: 10/Mar/22 12:24

Description

Currently, Spark SQL can throw Java exceptions from the timestampadd()/date_add()/dateadd() functions, for instance:

spark-sql> select timestampadd(YEAR, 1000000, timestamp'2022-03-09 01:02:03');
22/03/09 14:47:15 ERROR SparkSQLDriver: Failed in [select timestampadd(YEAR, 1000000, timestamp'2022-03-09 01:02:03')]
java.lang.ArithmeticException: long overflow
	at java.lang.Math.multiplyExact(Math.java:892) ~[?:1.8.0_292]
	at org.apache.spark.sql.catalyst.util.DateTimeUtils$.instantToMicros(DateTimeUtils.scala:505) ~[spark-catalyst_2.12-3.3.0-SNAPSHOT.jar:3.3.0-SNAPSHOT]
	at org.apache.spark.sql.catalyst.util.DateTimeUtils$.timestampAddMonths(DateTimeUtils.scala:724) ~[spark-catalyst_2.12-3.3.0-SNAPSHOT.jar:3.3.0-SNAPSHOT]
	at org.apache.spark.sql.catalyst.util.DateTimeUtils$.timestampAdd(DateTimeUtils.scala:1197) ~[spark-catalyst_2.12-3.3.0-SNAPSHOT.jar:3.3.0-SNAPSHOT]

That might confuse non-Scala/Java users. Need to wrap such kind of exception by Spark's exception using an error class.

Issue Links

• is a clone of: SPARK-37858 Throw Spark exceptions from AES functions
• links to: [Github] Pull Request #35787 (MaxGekk)
• links to: [Github] Pull Request #35787 (MaxGekk)