Consider adopting fastutil for deserialization of primitive collections
- Dominant language
- Java
- Stars
- 87
- Forks
- 73
- Avg merge
- 2h 50m
- Merged PRs (30d)
- 1
Description
Array and Maps of primitive types are common in Avro schema. Java collections do an acceptable generic job, but type-specific collections such as [fastutil](https://fastutil.di.unimi.it/) can save overhead in multiple ways:
1. No boxing/unboxing of numbers.
2. Primitive type-specific collections lower memory overhead of underlying representations.
3. Which is more pronounced in JVM heaps over 32gb due to 16bit object reference overhead.
fastutil is well respected and widely adopted for performance, so using it for Java representation of Avro data could provide a nice benefit.
Contributor guide
Research direction
The issue names no files, tests, or entry points. Start by locating the Java deserialization code for Avro arrays and maps, then determine the scope and compatibility requirements for adopting fastutil; done would require an agreed design and implementation plan.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- java
- Domain
- backend
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100