For each set of query predicates forming the basis for an HBase scan, a set of context rich, data "flavor" specific structures are constructed to abstract the complexities involved in assembling the various segments and fields of composite (scan start/stop) row keys under various relational and logical operator and various configurable composite key-field hashing, formatting, padding and other features. If an HBase partial row-key scan is possible based on the literal set, a {@link PartialRowKeyScanAssembler} is invoked using the current scan context resulting in a pair of composite start/stop row keys. These are then used in the HBase client scan API start and stop row.