Range Partitioning

Range partitioning is a type of relational database partitioning wherein the partition is based on a predefined range for a specific data field such as uniquely numbered IDs, dates or simple values like currency. A partitioning key column is assigned with a specific range, and when a data entry fits this range, it is assigned to this partition; otherwise it is placed in another partition where it fits.

In a range partitioned table, rows are distributed based on a "partitioning key" where the only requisite is whether or not the data falls within the range specification of the key. For example, if the partition key is a date column, and January 2015 is a partition, then all data containing values from January 1, 2015 to January 31, 2015 will be placed in this partition.

Range partitioning is quite useful for applications requiring high performance for both decision-support environments and online transaction processing (OLTP). This makes data segregation easy and access to each smaller partition is fast, however extensive knowledge as to the data partitioning is required in order to properly balance the load evenly across all partitions. In this scheme, many partitions are ordered, with each succeeding partition having a higher bound than the previous partition.

Characteristics of range partitioning:

• Each partition has an exclusive upper bound.
• Each partition has a non-inclusive lower bound, except for the very first partition.