[SPARK-10223] Add takeOrderedByKey function to extract top N records within each group

• Type: New Feature
• Status: Closed
• Priority: Minor
• Resolution: Won't Fix
• Component/s: PySpark
• Assignee: Unassigned
• Reporter: Ritesh Agrawal
• Shepherd: DB Tsai
• Created: 25/Aug/15 13:05
• Updated: 16/Oct/16 12:34
• Resolved: 16/Oct/16 12:34

Description

Currently PySpark has takeOrdered function that returns top N records. However often you want to extract top N records within each group. This can be easily implemented using combineByKey operation and using fixed size heap to capture top N within each group. A working solution can be found over [here](https://ragrawal.wordpress.com/2015/08/25/pyspark-top-n-records-in-each-group/)