[SPARK-35998] Make from_csv/to_csv to handle year-month intervals properly

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.2.0, 3.3.0
• Fix Version/s: 3.2.0
• Component/s: SQL
• Assignee: Kousuke Saruta
• Reporter: Kousuke Saruta
• Created: 02/Jul/21 19:36
• Updated: 05/Jul/21 10:11
• Resolved: 05/Jul/21 10:11

Description

from_csv throws exception if year-month interval types are given.

spark-sql> select from_csv("interval '1-2' year to month", "a interval year to month");
21/07/03 04:32:24 ERROR SparkSQLDriver: Failed in [select from_csv("interval '1-2' year to month", "a interval year to month")]
java.lang.Exception: Unsupported type: interval year to month
	at org.apache.spark.sql.errors.QueryExecutionErrors$.unsupportedTypeError(QueryExecutionErrors.scala:775)
	at org.apache.spark.sql.catalyst.csv.UnivocityParser.makeConverter(UnivocityParser.scala:224)
	at org.apache.spark.sql.catalyst.csv.UnivocityParser.$anonfun$valueConverters$1(UnivocityParser.scala:134)

Also, to_csv doesn't handle year-month interval types properly though any exception is thrown.
The result of to_csv for year-month interval types is not ANSI interval compliant form.

spark-sql> select to_csv(named_struct("a", interval '1-2' year to month));
14

The result above should be INTERVAL '1-2' YEAR TO MONTH.

Issue Links

• links to: [Github] Pull Request #33210 (sarutak)