[SPARK-10648] Spark-SQL JDBC fails to set a default precision and scale when they are not defined in an oracle schema.

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.5.0
• Fix Version/s: 1.4.2, 1.5.3, 1.6.0
• Component/s: SQL

Environment

using oracle 11g, ojdbc7.jar

• Target Version/s: 1.6.0
• Assignee: Travis Hegner
• Reporter: Travis Hegner
• Created: 16/Sep/15 20:25
• Updated: 04/Jan/16 14:30
• Resolved: 05/Nov/15 20:37

Description

Using oracle 11g as a datasource with ojdbc7.jar. When importing data into a scala app, I am getting an exception "Overflowed precision". Some times I would get the exception "Unscaled value too large for precision".

This issue likely affects older versions as well, but this was the version I verified it on.

I narrowed it down to the fact that the schema detection system was trying to set the precision to 0, and the scale to -127.

I have a proposed pull request to follow.

Issue Links

• is duplicated by: SPARK-10909 Spark sql jdbc fails for Oracle NUMBER type columns
• is related to: SPARK-10909 Spark sql jdbc fails for Oracle NUMBER type columns
• relates to: SPARK-10909 Spark sql jdbc fails for Oracle NUMBER type columns
• links to: [Github] Pull Request #8780 (travishegner)
• links to: [Github] Pull Request #9495 (travishegner)
• links to: [Github] Pull Request #9498 (yhuai)