Dividing the nn data sets into scan groups

Edit online

Estimate the sizes of the work data sets with suffix nn that are used by each scan group.

Procedure

Refer to the following examples to make these estimates:
Example 1

Assume that you specify DBA and DBB to be processed by scan group 01, and DBC to be processed by scan group 02. Also assume that each database requires MERGINnn as shown in the following worksheet.

Table 1. Worksheet 2 for Example 1

The values for IXKEY and CHECKREC are common across databases.
A. DB name i. MERGINnn ii. SORTEX01 iii. MERGI2nn iv. SORTE2nn v. IXKEY vi. SORTILnn vii. CHECKREC
DBA 1000
DBB 2000
DBC 5000

You can estimate the size of MERGIN01 and MERGIN02 as follows:

  • Size of MERGIN01= 1000 + 2000 = 3000 bytes
  • Size of MERGIN02= 5000 bytes
Example 2

In addition to the assumptions made in Example 1, assume that DBA has two data sets, and the size of data set 1 is 1 GB and that of data set 2 is 4 GB. Also assume that you specify the processing of each scan group as follows:

  • Scan group 01: Data set 1 of DBA, and DBB
  • Scan group 02: Data set 2 of DBA, and DBC

The MERGINnn for each data set of DBA requires the following size:

  • For data set 1: (1000 x 1G/(1G+4G) ) = 200 bytes
  • For data set 2: (1000 x 4G/(1G+4G) ) = 800 bytes

Then your estimates of the sizes of MERGIN01 and MERGIN02 will be as follows:

  • Size of MERGIN01 = 200 + 2000 = 2200 bytes
  • Size of MERGIN02 = 800 + 5000 = 5800 bytes

You can estimate the sizes of other work data sets with suffix nn in the same way as you did for MERGINnn. For SORTEX01, which is used per scan job step, however, you need to estimate the size as follows:

  • If you run all of SCAN processes in one job, add all SORTEX01 columns.
  • If you run the SCAN processes in multiple job steps, divide worksheet 2 by scan jobs and calculate SORTEX01, MERGINnn, and the other work data sets by each worksheet.