[SPARK-14319] Speed up group-by aggregates

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Done
• Fix Version/s: 2.0.0
• Component/s: SQL
• Assignee: Unassigned
• Reporter: Sameer Agarwal
• Created: 31/Mar/16 23:03
• Updated: 12/Dec/22 18:10
• Resolved: 03/Nov/16 03:04

Description

Aggregates with key in SparkSQL are almost 30x slower than aggregates with key. This master JIRA tracks our attempts to optimize them.

Sub-Tasks

• Benchmark Vectorized HashMap for GroupBy Aggregates: Resolved
• Make ColumnarBatch.Row mutable: Resolved
• Generate AggregateHashMap class during TungstenAggregate codegen: Resolved
• Speed up TungstenAggregate w/ keys using AggregateHashMap: Resolved
• Use/benchmark a better hash in AggregateHashMap: Resolved
• Support all datatypes to use VectorizedHashmap in TungstenAggregate: Resolved
• Disable vectorized hashmap for wide schemas and high-precision decimals: Resolved