[SPARK-3056] Sort-based Aggregation

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.5.0
• Component/s: SQL
• Assignee: Yin Huai
• Reporter: Cheng Hao
• Created: 15/Aug/14 00:54
• Updated: 28/Jul/15 20:08
• Resolved: 22/Jul/15 06:27

Description

Currently, SparkSQL only support the hash-based aggregation, which may cause OOM if too many identical keys in the input tuples.

Issue Links

• is blocked by: SPARK-2926 Add MR-style (merge-sort) SortShuffleReader for sort-based shuffle
• is duplicated by: SPARK-2873 Support disk spilling in Spark SQL aggregation
• links to: [Github] Pull Request #7458 (yhuai)