File sizes required for the execution of the statistics analysis utility (pdstedit)

The following table shows the formulas for determining the file sizes required for the execution of the statistics analysis utility (pdstedit):

File type		Formula (bytes)
Temporary work file	Statistical information about system activity	4,096 x collection count^# x 2
	Statistical information about system activity per server	4,096 x collection count^# x number of servers x 2
	Statistical information about UAPs	872 x number of UAPs to be executed or number of transactions to be executed x 2
	Statistical information about SQL	512 x number of SQLs to be executed x 2
	Statistical information about SQL static optimization	512 x SQL object cache mishit count
	Statistical information about SQL dynamic optimization	49,624 x number of SELECT statements issued by OPEN or EXECUTE (including INSERT SELECT)
	Statistical information about SQL object execution	512 x number of SQLs to be executed x number of servers
	Statistical information about SQL object transfer	256 x number of SQLs to be executed x number of servers
	Statistical information about the history of SQL statements	(1,024 + average SQL length) x number of SQLs to be executed
	Statistical information about CONNECT/DISCONNECT	256 x number of CONNECTs and DISCONNECTs
	Statistical information about global buffer	512 x number of synchronization points x 2
	Statistical information about database manipulation for HiRDB files	512 x number of synchronization points x 2
	Statistical information about deferred write processing	512 x number of deferred write operations x 2
	Statistical information about indexes (input: STJ)	128 x number of synchronization points x 2
	Statistical information about indexes (input: FJ)	128 x number of page splits x 2
Work file for sorting	Work area for sorting the above types of temporary work files for analysis	Maximum size of the temporary work files that will be needed for analysis
DAT- format file	Statistical information about system activity	3,404 x collection count^#
	Statistical information about system activity per server	3,262 x collection count^# x number of servers
	Statistical information about UAPs	1,991 x number of UAPs to be executed or number of transactions to be executed
	Statistical information about SQL	447 x number of SQLs to be executed
	Statistical information about SQL static optimization	646 x SQL object cache mishit count
	Statistical information about SQL dynamic optimization	380 x number of SELECT statements issued by OPEN or EXECUTE (including INSERT SELECT)
	Statistical information about SQL object execution	520 x number of SQLs to be executed x number of servers
	Statistical information about SQL object transfer	389 x number of SQLs to be executed x number of servers
	Statistical information about the history of SQL statements	(240 + average SQL length) x number of SQLs to be executed
	Statistical information about CONNECT/DISCONNECT	278 x number of CONNECTs and DISCONNECTs
	Statistical information about global buffer	600 x number of synchronization points
	Statistical information about database manipulation for HiRDB files	356 x number of synchronization points
	Statistical information about deferred write processing	300 x number of deferred write operations

#: Collection count = (pdstend command input time - pdstbegin command input time) interval specified with the -m option

Organization of this subsection: (1) Approximate size of a temporary work file; (2) Approximate size of a work file for sorting; (3) Approximate size of a DAT format file

(1) Approximate size of a temporary work file

You can use the formula shown below to obtain an approximation of the size of a temporary work file that is created during execution of the statistics analysis utility. This formula assumes that all information in the statistics log file that is input to the utility is analyzed.

Size of temporary work file = size of statistics log file x a (bytes)

a

Use one of the values listed below, depending on the type of statistical information contained in the statistics log file. If the statistics log file contains more than one type of statistical information, use the largest value for a.

Type of statistical information	Value for a
System activity statistical information	3.4
SQL dynamic optimization information	135.6
Other	2

About errors

If the statistics log file contains multiple types of statistical information and one of them is SQL dynamic optimization information, the difference between the approximated value and the actual value might be large depending on the ratio of the SQL dynamic optimization information contained in the file. If you want to reduce this difference, make a separate calculation for SQL dynamic optimization information.

You can use the following formula to obtain the temporary work file size for SQL dynamic optimization information:

Size of temporary work file for SQL dynamic optimization information = number of SQL dynamic optimization information items x 49,624 (bytes)

You can obtain the number of SQL dynamic optimization information items from the input log file summary information that is output when you execute the statistics analysis utility. For details about the input log file summary information, see Statistics analysis utility (pdstedit) in the manual HiRDB Version 9 Command Reference.

(2) Approximate size of a work file for sorting

You can use the formula shown below to obtain an approximation of the size of a work file for sorting that is created during execution of the statistics analysis utility. This formula assumes that all information in the statistics log file that is input to the utility is analyzed.

Size of work file for sorting = size of statistics log file x b (bytes)

b

Type of statistical information	Value for b
System activity statistical information	1.7
Other	1

19.1.3 File sizes required for the execution of the statistics analysis utility (pdstedit)

(1) Approximate size of a temporary work file

(2) Approximate size of a work file for sorting

(3) Approximate size of a DAT format file