Using the Central Component › Data Mining › Define the Data Mining Scope › Data Mining Selection Panel

Data Mining Selection Panel

Use this panel to do the following:

1. In the OPTION field, select the Data Mining Scope by entering one of the following option numbers:
   1. Select measurements of all job steps stored in Performance Management Assistant.
   2. Select measurements of only the job steps that belong to the TOP Scope. For details about the TOP Scope, see Use the TOP scope Facility.
   3. Select measurements of the first nnn job steps in alphabetical order where nnn is a value between 1 and 999.
2. To optionally filter the data that is mined, enter specific search criteria as follows:

   Jobname

   The job name of relevant job steps using underscore as a wildcard placeholder.

   PGM Name

   The name of relevant programs using underscore as a wildcard placeholder.

   From Date

   The oldest creation date of relevant measurements.

   Importance

   The lower limit of job importance (see job file).

   Base Date

   Defines a different end date to weight the current frequency counter for the estimated annual frequency calculation (optional).

   To calculate the annual frequency, the current frequency is set in relation to the dates of the first job step execution that Performance Management Assistant received data from and the base date.

   The difference between the two dates and the current job step executions in that time frame is the base for the estimate of the annual forecast calculation.

   The default is the actual date.

   Note: This date is not considered if Annual Freq. = N is used.

   Annual Freq.

   Controls whether the job step frequency (and therefore the CPU time consumption) is calculated as an annual forecast or if the frequency is taken as it is and the CPU time consumption is calculated based on this current frequency.

   Y = Frequency and CPU time are calculated as an annual forecast.

   N = Frequency and CPU time are calculated based on the current job step frequency.

   The restriction criteria are combined with logical AND

3. The following calculation methods are available for the calculation of annual CPU time:

   AVG

   The base is the average CPU time of the job step.

   LAST

   The base is the CPU time of the last (most recent) job step execution.

   MEAS

   The base is the CPU time of the measured job step.

   If the MEAS method is selected and the measurement does not fit any entry in the job history list, the annual CPU time is calculated as follows:

                measurement CPU time * measurement elapsed time * frequency
   Annual CPU time =---------------------------------------------------
                           job step average elapsed time
   

4. Press Enter to perform the search. The Data Mining Menu displays. If the mining process requires more than 5 seconds, in progress updates are displayed until it is completed.
5. If the message 4000 rows exceeded displays, you should perform step 2 (again) to limit the search.