STATISTICS command with SAVE subcommand

Usage

You can also use the converted SAVE data sets to generate CSV (comma-separated value) input-data. This CSV data can then be transferred to workstations and imported into spreadsheets to improve DB2 performance analysis using graphical representations or pivot tables. For more information refer to Reporting User's Guide.

Usage notes

• VSAM data sets cannot be concatenated.
• For information on migration refer to Report Reference.

Syntax of the SAVE subcommand

Subcommand options

The syntax diagram shows the options that are available with this subcommand. See Subcommand options for comprehensive descriptions of these options. The following list gives additional or specific descriptions of selected options, where appropriate.

DDNAME

Specifies the ddname where the save data is written. The default ddname is STSAVDD. You can specify any valid ddname provided that your JCL contains a valid DD statement for it.

Without CONVERT option:

When using SAVE, your JCL can also include a valid DD statement for JSSRSDD, so that the related job summary information can be saved.

Note: Do not specify DUMMY in the JCL for either STSAVDD or JSSRSDD. When you are not using SAVE, omit these ddnames from your JCL.

The VSAM data set defined by the default ddname must already exist when you run OMEGAMON for DB2 PE. See the Report Reference for a description of VSAM data sets. Either specify an existing data set from a previous OMEGAMON for DB2 PE run (when restoring data), or specify a new data set allocated using the IDCAMS DEFINE CLUSTER function.

Note: If ddname is assigned to a non-VSAM file, you receive an error message and the job terminates.

With CONVERT option:

The ddname needs to be assigned to a physical sequential data set. This data set can be used for a subsequent load to Statistics SAVE tables.

Note: If ddname is assigned to a nonsequential data set, you receive an error message and the job terminates.

DSETSTAT

Include buffer pool data set statistics. This option is ignored if option INTERVAL(X) with X>0 has been specified. Statistics interval processing with intervals > 0 can only be applied to IFCIDs 1, 2, 225, and 369 that are externalized by DB2 every minute. IFCID 199 externalization is controlled by system parameter STATIME and can range from 1 to 60 minutes. Interval metrics for buffer pool data sets can be derived from Statistics TRACE or FILE data created with DSETSTAT.

CONVERT

Specifies that converted reduced data is written directly to a sequential data set.

It is recommended to specify this option if you want to create high amount of loadable reduced data in a sequential data set. It avoids a temporary VSAM data set to be used as a SAVE data set. This options causes OMEGAMON for DB2 PE to write converted reduced data directly to a sequential data set in a single step. The resulting output is loadable to the PDB tables. The user may experience performance improvements compared to the default path with a separate SAVE step and the subsequent convert of saved data by the save-file utility.

Note: Not every big trace input results in big reduced data and small trace input in small reduced data. This option becomes effective when the reduction results in high amount of reduced data. For example, it depends on the amount of different criteria that results due to your filtering. It is possible that you experience more performance improvement with a small input trace that results in many different criteria than with a large trace input with only a few different criteria.

This option can generally be used as it is neutral to the performance when processing only a few reduced records, but becomes effective when processing a high amount of reduced data. However, do not use this option if you want to RESTORE and REPORT saved data.