Usually, when an application adds a record to a partitioned file, the value of the primary key determines which disk volume stores the record. Over time, however, this event can lead to an uneven distribution of records across the various partitions in a file. Ultimately, this can cause an application outage due to a file full condition, or it can compromise application performance due to excessive disk input/output (I/O). Rearranging the partitions in the file solves the problem of uneven record distribution, but to do this, users must specify the lowest primary key value for the
records to be stored in each partition. The choice of these key values is critical to the successful redistribution of records within the file’s partitions. Organizations need a solution that provides the exact information necessary to determine the key values that will lead to the desired distribution of records in a partitioned file quickly and accurately.