Creating Group ID from SUM Limit - response (2) by saraider99

I was thinking I may be able to get away with simply distributing the records evenly based on an average or other distribution key. If I had data such as below I would want to distribute them into buckets somewhat evenly.
PROD A - 50
PROD B - 55
PROD C - 1000
PROD D - 25
PROD E - 32
PROD F - 800
One possible result
GRP 1 - (PROD A & D)
GRP 2 - (PROD B & E)
GRP 3 - (PROD C)
GRP 4 - (PROD F)
Is that possible?