Administering SQR for z/OS
This section provides an overview of SQR on z/OS and discusses how to run SQRs outside of Process Scheduler and specify the input and the output files.
The SQR product is available for z/OS server platforms. The z/OS version of SQR is compatible with your existing SQR reports that you currently run from your client machines. The ability to run SQRs on the mainframe means a significant performance enhancement for SQR execution. All SQL is dynamic, therefore no precompiling is necessary for running SQRs.
You can execute SQRs on z/OS by either by using PeopleSoft Process Scheduler or submitting them as traditional z/OS jobs. Process Scheduler dynamically generates SQR JCL.
Be aware that some application processes are designed to run only through Process Scheduler.
SQR for z/OS is delivered with the PeopleTools installation. The PeopleTools DB2 UDB for z/OS installation guide includes instructions for performing the installation of SQR on the z/OS server. SQR must be installed prior to running PeopleSoft SQRs on the z/OS server.
Allow at least 12 cylinders of 3390 DASD or equivalent disk space to complete the installation.
These run time parameters are supplied to SQRs initiated by the Process Scheduler.
PROCESS_INSTANCE
OPRID
RUN_CNTL_ID (required by specific applications only, such as Financials)
When an SQR request is submitted by Process Scheduler, run time parameters are obtained from a Process Scheduler table and dynamically inserted into the generated JCL.
For an SQR submitted as a traditional z/OS batch job, two run time parameters are required, but it is not necessary to pass any specific values. It is only necessary to include a blank line for each parameter in the JCL specified in the SYSIN DD. The appropriate segment of the JCL is shown below:
//SQRNAME EXEC SQRPROC,SQRID=SQRNAME
//SQR.SYSIN DD *
/*
Sample SQR JCL is provided in HLQ.PPVVV.JCLLIB(SQRSAMP).
Input and output files are required by SQR when using commands such as OPEN and NEW-REPORT. Remember to consult the appropriate PeopleSoft application documentation for important application specific information concerning SQR for z/OS. For instance, Accounts Payable naming conventions used to build DD names are discussed in the PeopleSoft Payables product documentation.
There are two ways to specify input and output files in SQR for z/OS:
Add DD statements to the JCL.
This method allows you to maintain a common set of SQR that can execute on any operating system by requiring the modification of Process Scheduler Shell JCL and/or z/OS Batch JCL.
Add a DSN style filename to the SQR.
This method alleviates the need to modify JCL but creates z/OS operating system specific SQR.
Use the first method if your SQR should execute on any operating system. Use the second method if your SQR executes only on the z/OS operating system.
Adding DD Statements to the JCL
You may specify a file name for commands such as OPEN and NEW-REPORT using DOS or UNIX file naming conventions. The SQR for z/OS documentation states that SQR will use up to 8 alpha-numeric characters preceding the file extension as a DD name in JCL.
SQR for z/OS does not find 8 alpha-numeric characters as documented. To get around this problem, the FILEPREFIX and FILESUFFIX environment variables in $PSHLQ$.SQRINC(SETENV) are used when coding filenames in SQR. The example below shows that SETENV does not contain values for FILEPREFIX and FILESUFFIX when running on the z/OS operating system:
! File prefixes and suffix
!
#ifdef NT
#define FILEPREFIX C:\TEMP\
#define FILESUFFIX
#endif
!
#ifdef MVS
#define FILEPREFIX
#define FILESUFFIX
#endif
!
#ifdef UNIX
#define FILEPREFIX /usr/tmp/
#define FILESUFFIX
#endif
!
This coding standard enables DOS or UNIX file naming conventions to be used with the OPEN or NEW-REPORT SQR commands. The root portion of the file name is used as a JCL DD name. The Process Scheduler Shell JCL or z/OS Batch JCL must contain this DD name. Each file used for input or output requires a separate DD statement in the JCL.
In the following example, SQR for z/OS uses VIEWTBL as the DD name in JCL:
let $outputfile = ‘{FILEPREFIX}VIEWTBL{FILESUFFIX}’
open $outputfile as 1 for-writing record=132
The DD statement in the execution JCL requires the same DD name:
//VIEWTBL DD DSN=&PSHLQ..OUTFILES(VIEWTBL),DISP=SHRThe data set name specified in the JCL may be either sequential or partitioned dataset. If the SQR requires multiple input or output files, you must add a separate DD statement to the JCL for each file.
While modifications to Process Scheduler Shell JCL or z/OS Batch JCL are required using this method, the resulting SQR is not operating system specific.
Note: DD names must reference either sequential datasets or separate partitioned datasets due to the z/OS restriction on writing to more than one PDS member simultaneously.
Note: The system overrides the FILEPREFIX and FILESUFFIX variables with the value specified in setfile.sqc, when running the SQR reports through the Process Scheduler with an output type of Web. The FILEPREFIX variable will be substituted with the log_output folder path of the Process Scheduler domain (<DOMAIN>/log_output/SQR_<process_name>_<prcs_instance>/). The FILESUFFIX variable will be substituted with a blank value. When there is a requirement to generate a static file or provide a static file input to the SQR report for an output type of Web, these variables should be redefined in the SQR file itself (rather than in the SQC file) to any suitable path and extension.
Adding a DSN Style Filename to the SQR
Filenames are preceded by DSN: (dataset name). For example:
OPEN ‘DSN: $PSHLQ$.SQR.DAT(VIEWTBL)’ FOR-READING RECORD=133
NEW-REPORT ‘DSN: $PSHLQ$.SQR.DAT(VIEWTBL)’
Modifications to Process Scheduler Shell JCL or z/OS Batch JCL are not required when using the above option. However, this option results in operating system specific SQR that executes only on the z/OS operating system.
The SQR language supports a DECLARE PRINTER command so that reports can be directed to HPLASERJET and POSTSCRIPT type printers. However, if the printer is not mainframe-connected, the TYPE=LINEPRINTER option is required.
The TYPE=LINEPRINTER specification produces a basic text type report which can be redirected to a system printer. Normally, print output lines don't exceed 124 print positions.
SQRs containing the SETUP02 statement (refers to a member in the SQC library) allow print lines up to 177 positions. If SQROUT DD prints to SYSOUT, then supply a DCB override for SYSOUT and set the LRECL to 178, this is given in SQRSAMP JCL under your JCLLIB PDS. It would be a good practice to have DCB override with LRECL=178 for SQROUT DD, as this fits both landscape and portrait mode.
Formatted reports cannot be downloaded, then printed from a workstation connected printer. Only selected SQRs utilize the DECLARE PRINTER feature.
Note: A “formatted” report is one produced with the TYPE=HPLASERJET /POSTSCRIPT specification.
Many customers customize SQRs to tailor information to unique business requirements.
Another reason to customize SQRs is due to the way SQR formats reports when the TYPE=LINEPRINTER option is used. Column header and data field alignment may not be optimal, so modifications may be required.