The size of the statistics log that is output differs depending on the statistical information. The size of the statistics log for each type of statistical information is described starting in (1) as follows. For the pd_stj_file_size operand, specify a value that is equal to or greater than the combined total of the statistics log sizes determined in (1) and beyond.
This will depend on the options specified in the pdstbegin operand or pdstbegin command.
If the total number of UAPs that will actually be executed is known, use the formula above. If the number of UAPs that will be executed during statistical information collection is not known, that number can be determined with the following formula:
For example, if 10 UAPs are executed in 30 minutes and statistical information is collected for a period of 60 minutes, the UAP count will be 10 60
30 = 20.
If the total number of SQLs that will actually be executed is known, use the formula above. If the number of SQLs that will be executed during statistical information collection is not known, that number can be determined with the following formula:
#: If a stored procedure or stored function will be executed, the SQL statements in the procedure or function must also be counted.
If it is possible to determine the actual number of SQL objects to be created, this formula can be used to make that determination. If the number of SQL objects to be created during statistical information collection is not known, the number of SQL objects to be created can be determined using the following formula:
Number of SQL objects to be created = average-number-of-SQL-statements-to-be-executed-in-UAP SQL-object-cache-miss-rate
(number-of-UAP-executions-per-unit-of-time
statistical-information-collection-time
unit-of-time)
The SQL object cache miss rate can be determined from the statistical information related to the system.
HiRDB single server configuration
HiRDB parallel server configuration
If the total number of SQLs that will actually be executed is known, use the computation formula above. If the number of SQLs that will be executed during statistical information collection is not known, that number can be determined with the following formula:
Number of SQLs# = average-number-of-SQL-statements-executed-per-UAP (number-of-UAPs-to-be-executed-per-unit-time
(statistical-information-collection-period
unit-time)
#: If a stored procedure or stored function will be executed, the SQL statements in the procedure or function must also be counted.
If it is possible to determine the cumulative number of SQL statements to be executed, this formula can be used to make that determination. If the number of SQL statements to be executed during statistical information collection is not known, the number of SQL statements can be determined using the following formula:
Number of SQL statements# = average-number-of-SQL-statements-to-be-executed-in-UAP (number-of-UAP-executions-per-unit-of-time
statistical-information-collection-time
unit-of-time)
#: If a stored procedure or stored function is to be executed, the SQL statements described in the procedure or function must also be included in the count.
This differs depending on the option specified for the pdstbegin operand or pdstbegin command.