I have a database of 164,000 objects. The keys are not very evenly distributed, so I\'m using md5 to hash each key, and I create buckets based on the first two characters o
