Sylvain Wallez wrote:
The org.apache.hadoop.record package contains some very interesting stuff. However, I can't find any use of it within Hadoop itself (nor in Nutch). Is it because it's intended for application-data use only?
No, we just haven't had a chance to start using it. The long-term plan is to use it widely.
Its cross-platform abilities make me think it could be an interesting base to build a Sawzall-like thing [1] and I'm thinking to use the Rhino JavaScript interpreter for this. WDYT?
Sounds reasonable to me! Doug
