REORG phase for a single-phase reorganization
- Writes discarded rows from SELECT or DELETE processing to the SYSARC data set (if specified)
- Builds or keeps the compression dictionary and compresses the rows
- If required, adds partitions to a partition-by-growth table space (SHRLEVEL NONE and SHRLEVEL REFERENCE only)
- Redefines the VSAM data sets (including the staging data sets for SHRLEVEL REFERENCE or SHRLEVEL CHANGE) when the value of the REDEFINE command or installation option is YES
- Checks for duplicate key values in a unique index
- Concurrently loads multiple partitions
- Concurrently loads the table space and builds data-sorting indexes
- Concurrently sorts and builds non-data-sorting indexes
- Concurrently builds multiple, non-data-sorting indexes
- Collects statistics while loading table spaces and building indexes
- Produces image copies either while reloading data or after reloading rows
If you have a large number of nonpartitioned indexes, consider using a two-phase reorganization to have REORG PLUS concurrently reload the table space and build the nonpartitioned indexes. Otherwise, a single-phase reorganization provides significant CPU and elapsed time savings over a two-phase reorganization because REORG PLUS does not have to read from the SYSREC data set (for a table space reorganization) or the SYSUT1 data set (for an index reorganization).
For a single-phase reorganization, the considerations in the following table apply to the SYSREC data set for a table space reorganization or the SYSUT1 data set for an index reorganization and depend on the SHRLEVEL that you specify.
Specified SHRLEVEL | Table space reorganization | Index reorganization | Effect on restartability |
|---|---|---|---|
SHRLEVEL NONE (the default) | You can omit SYSREC data set for an additional performance gain. | You can omit SYSUT1 data set for an additional performance gain. | If you omit the data set, your job might not be restartable. |
SHRLEVEL REFERENCE | REORG PLUS does not use SYSREC data set, even if you specify it. | REORG PLUS does not use SYSUT1 data set, even if you specify it. | The job is restartable due to the non-destructive nature of this type of reorganization. |
SHRLEVEL CHANGE | REORG PLUS does not use SYSREC data set, even if you specify it. | REORG PLUS does not use SYSUT1 data set, even if you specify it. | The job is not restartable until the UTILTERM phase begins. Restarting the job before the UTILTERM phase begins is unnecessary because the original data sets are unaffected. |
Except for the option of omitting the SYSREC or SYSUT1 data set, all of the performance considerations for the REORG phase are the same as those in the UNLOAD phase (UNLOAD-phase) and the RELOAD phase (RELOAD-phase). The following figures show the objects that REORG PLUS uses in the REORG phase.
REORG phase: SHRLEVEL NONE or SHRLEVEL REFERENCE UNLOADONLY table space reorganization
REORG phase: SHRLEVEL REFERENCE table space reorganization
REORG phase: SHRLEVEL NONE or SHRLEVEL REFERENCE UNLOADONLY index reorganization
REORG phase: SHRLEVEL REFERENCE index reorganization
