keyrange

Keyrange is a term used in information systems to denote a contiguous interval in the key space that bounds a set of keys. It is commonly used to describe data partitions, range scans, or routing boundaries in distributed storage and databases. A keyrange is typically defined by a lower bound and an upper bound, with optional flags to indicate whether the endpoints are inclusive or exclusive. Depending on the system, keys may be numeric or lexicographically ordered strings, and the exact ordering rules determine how boundaries are interpreted.

In practice, keyranges enable partitioning and pruning: a shard, node, or region can be responsible for all records whose keys lie within its keyrange, allowing queries or scans to skip irrelevant data. They also support data migration and backup by range, where ranges are moved or backed up independently.

For example, in a system with keys 0 through 9999, a shard might handle the keyrange [0, 1999], another [2000, 4999], and so on. In a document store using string keys, a keyrange could be ["user:0000", "user:0999"].

Implementation considerations include representing a KeyRange as a simple data structure or object with methods such as contains(key), intersects(other), and isEmpty(). Care is needed to be explicit about inclusive versus exclusive bounds and key encoding, as different systems use different conventions.

Related concepts include range partitioning, shard key, and range query. A keyrange is a descriptor of the data scope used by queries and routing decisions, rather than a query itself.