Estimating the sizes of the work data sets for EPSCHK=YES (SORTILnn)

Follow these instructions to estimate the sizes of the data sets that are needed when you specify EPSCHK=YES. The following processes are required only for HALDBs that have logical or index relationships.

Estimate the size of SORTILnn.

Records are written in SORTILnn during the SCAN process of TYPE=SCAN. To calculate the value for SORTILnn in worksheet 2, complete the following steps:

1. Fill in column J in worksheet 1.

   Column J: Total number of pointers in EPS

   Put the following value in column J:

   • If column B is PSINDEX, the same number as in column I.
   • If column B is PHDAM or PHIDAM, see the SEGMENT AND POINTER STATISTICS part in the Database Statistics report. Take the values in TOTAL under USED from the ELP and the ELC rows, and add up the values for all segments in the database. Put the sum in column J.

2. Calculate SORTILnn in worksheet 2 by the following formula:
   • If IXKEYCHK=YES is specified for PSINDEX:

        SORTILnn = (56 + key length) x J Bytes

     Where key length is the value shown in DB INDEX KEYLENGTH-1 field of the Scan of Index Database report plus 1.
   • If IXKEYCHK= NO is specified:

        SORTILnn = 44 x J Bytes