[SPARK-42442] Use spark.sql.timestampType for data source inference

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.4.0
• Fix Version/s: 3.5.0
• Component/s: SQL
• Assignee: Gengliang Wang
• Reporter: Gengliang Wang
• Created: 14/Feb/23 23:27
• Updated: 15/Feb/23 11:39
• Resolved: 15/Feb/23 11:39

Description

With the configuration `spark.sql.timestampType`,  TIMESTAMP in Spark is a user-specified alias associated with one of the TIMESTAMP_LTZ and TIMESTAMP_NTZ variations. This is quite complicated to Spark users.

There is another option `spark.sql.sources.timestampNTZTypeInference.enabled` for schema inference. I would like to introduce it in https://github.com/apache/spark/pull/40005 but having two flags seems too much. After thoughts, I decide to merge `spark.sql.sources.timestampNTZTypeInference.enabled` into `spark.sql.timestampType` and let  `spark.sql.timestampType` control the schema inference behavior.

We can have followups to add data source options "inferTimestampNTZType" for CSV/JSON/partiton column like JDBC data source did.

Issue Links

• links to: [Github] Pull Request #40022 (gengliangwang)