ASAP 2.7 Extension Manual

Configuring the ASAP Extension

HP NonStop ASAP Extension Manual—425265-004

2-17

Formatting Metric Output

ASAPX DataItems other than text are stored as 64-bit floating point numbers, and all

metrics are initially computed as 64-bit floating values. The EDL TYPEDATA and

FORMAT properties control how ASAPX treats the data once a metric is computed. If

TYPEDATA is set to something other than REAL64, ASAPX translates the data to the

requested type before storing it in the APP record. If the value is too large to fit in the

requested type or to fit in the specified FORMAT, the field is set to all 9s.

You can set each attribute as TYPEDATA REAL64 and use the FORMAT statement to

control output. For example, FORMAT "I2" outputs a 2-digit integer if TYPEDATA is

REAL64, and FORMAT "F6.2" outputs a decimal value with 2 decimal places.

In versions of ASAPX prior to version 2.4 you were forced to use TYPEDATA REAL64

if you wanted to output a decimal value. You can now define decimal output for

TYPEDATA values of INT, INT32 or INT64. When ASAPX finds a decimal FORMAT

statement with a TYPEDATA of an integer type, it stores the value in the integer in

implied decimal format. For example, if TYPEDATA is INT64, FORMAT is "F6.2" and

the computed metric value is 123.45, ASAPX stores the value in the APP record as

12345.

Be careful using implied decimals with TYPEDATA values of INT and INT32.

Seemingly small numbers can cause an overflow. For example, if TYPEDATA is INT

and FORMAT is "F6.4," a number as small as 3.2767 overflows because ASAPX shifts

the decimal point 4 places to create 32767. INT64 is the recommended TYPEDATA for

implied decimals.

Some database analysis tools (Enform for example) cannot read floating point

numbers from ASAP records. To use those tools, define all user-created numeric

metrics as TYPEDATA INT64. Change ASAPX built-in metrics defined as REAL64 to

INT64.