Currently the IndexStrategy is responsible for generating single keys and query ranges. DynamoDB and Cassandra require the partition portion of the key be separate from the sort key.

This issue should lend itself to a follow on issue that should also give the ability to appropriately track the portions of the key for good data density informed input splits for analytics. A continuous histogram should be maintained only on the isolated portion of the key that is meant to be well-sorted and range scanned (the SFC). A separate discrete histogram should be maintained for the partition/hash portion.