The LOAD utility is used to accomplish bulk inserts to DB2 tables. It can add rows to a table, retaining the current data, or it can replace existing rows with the new data. The biggest LOAD performance considerations center on optimizing the running of the LOAD utility itself. Of course, inasmuch as LOAD adds data to DB2 tables, you need to be sure that you gather updated statistics by running either inline statistics or by following up each LOAD with a RUNSTATS. Failing to do so results in outdated statistics in the DB2 Catalog increasing the potential for improper access paths and poor application performance.
With LOAD utility there are lot of advantages. For instance, You can do mass deletes by specifying input data set DD DUMMY.
How to restart LOAD utility JCL
You can restart the JCL, by specifying UTPROC=RESTART in the JCL. When space abend (SB37) occurs with input dataset, just increse SPACE in I/p data set and RESTART the jcl.
If space abend occurs for O/P data set, you need to contact admin to increase O/P space.
Two methods for LOAD JCL
There are two distinct methods regarding the use of the LOAD utility.
- The first and generally recommended philosophy takes more time to implement but is easier to support. It requires the reservation of sufficient disk space to catalog the LOAD work data sets in case the LOAD job abends. The work data sets for the LOAD job are allocated for the DDNAMEs SORTOUT, SYSUT1, SYSERR, and SYSMAP with DISP=(MOD,DELETE,CATLG). This enables the data sets to be allocated as new for the initial running of the REORG job. If the job abends, it catalogs the data sets in case they can be used in a restart. After the step completes successfully, the data sets are deleted. The space for these data sets must be planned and available before the LOAD job runs.
- Second approach is you can create JCL with restart option. This helps when job fails for an abend, you can restart the job.