[SPARK-3933] Optimize decimal type in Spark SQL for those with small precision

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Component/s: SQL
• Assignee: Matei Alexandru Zaharia
• Reporter: Matei Alexandru Zaharia
• Created: 13/Oct/14 17:56
• Updated: 02/Nov/14 02:31
• Resolved: 02/Nov/14 02:31

Description

With fixed-precision decimals, many decimal values will fit in a Long, so we can use a Decimal class with a mutable Long field to represent the unscaled value, rather than allocating a BigDecimal. We can then do some operations directly on these Long fields.

Issue Links

• links to: [Github] Pull Request #2983 (mateiz)