[MAPREDUCE-1977] [Rumen] Do Not Store CDFs in Output JSON Files

• Type: Improvement
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: tools/rumen
• Assignee: Amar Kamat
• Reporter: Ranjit Mathew
• Created: 28/Jul/10 10:00
• Updated: 28/Jul/10 10:01

Description

Per-Job Cumulative Distribution Functions (CDFs) stored in JSON files emitted by Rumen are redundant and waste space (~30%).
They can easily be re-computed upon loading or just-in-time upon request.