scoobi
scoobi copied to clipboard
Implement skewed join
We have hit some performance issues when performing relational joins because of key skew. Pig and hive have implemented optimizations for skewed joins, and it would be nice for something similar to exist in scoobi.
Pig - http://ofps.oreilly.com/titles/9781449302641/advanced_pig_latin.html#skew_join Hive - https://cwiki.apache.org/confluence/display/Hive/Skewed+Join+Optimization