| CDB/RePart evenly distributes
the number of rows by creating a workfile with rows written
in clustering sequence, reads the workfile, defines a new
index and new partitions and re-writes both the new index
and new partitions. All of this is accomplished without
a DROP and recreate of the tablespace.
In addition, CDB/RePart concurrently
reorganizes output tablespaces and indexes, provides a concurrent
Image Copy of the new tablespace, optionally gathers and
updates runstats, and generates DDL for new partitioned
tablespaces with computed limitkey values.
CDB/RePart can also update
the DB2 catalog and directory, as well as issue a DB2-MODIFY
command to clear all records for the tablespace just processed.
Users benefit from increased DB2 system performance, data
distribution optimization, and no surprise outages due to
"overstuffed" partitions. |