*MINVAL/*MAXVAL - Evaluate the Minimum/Maximum

*MINVAL

([(IR=result-format/length)] operand,...)
*MAXVAL

Format/length: Format and length may be specified explicitly using the IR clause or evaluated automatically using the Format/Length Conversion Rule Tables below.

This document covers the following topics:

Function

The Natural system function *MINVAL/*MAXVAL evaluates the minimum/maximum value of all given operand values. The result is always a scalar value. If an array is specified as operand, the minimum/maximum of all array fields is evaluated.

When using alphanumerical or binary data as an argument, if the data is the same (for example, *MINVAL('AB','AB')), then the result is the argument with the smallest/largest length value.

Restrictions

When using the system function *MINVAL/*MAXVAL, the following restrictions apply:

  • *MINVAL/*MAXVAL must not be used where a target variable is expected.

  • You may not nest *MINVAL/*MAXVAL in a system function.

Syntax Description

Operand Definition Table:

Operand Possible Structure Possible Formats Referencing Permitted Dynamic Definition
operand C S A G   A U N P I F B D T         yes no

Syntax Element Description:

Syntax Element Description
*MINVAL Evaluates the minimum value of all given operand values.
*MAXVAL Evaluates the maximum value of all given operand values.
operand The operand(s) whose minimum/maximum values are to be evaluated by the *MINVAL/*MAXVAL system function.
result-format-length Intermediate Result clause for explicit specification of the resulting format/length. See IR Clause below.

IR Clause

The IR (Intermediate Result) clause may be used in order to specify explicitly the result-format/length of the whole *MINVAL/*MAXVAL system function. 

IR=result-format/length
IR=

format-length

(

A
U
B

) DYNAMIC

For an assortment of valid result-format/lengths, refer to the Format/Length Conversion Rule Tables below.

Syntax Element Description:

Syntax Element Description
format-length The compiler tries to determine the resulting format/length of the whole function. If the compiler cannot determine a format/length in a way that no loss of precision is guaranteed, the format-length must be set by the programmer using the IR operand extension.
A, U or B Format: Alphanumeric, Unicode or Binary for dynamic variable.
DYNAMIC Instead of specifying a fixed format/length, you may specify an alphanumeric, Unicode or binary format with dynamic length.

Example: 

DEFINE DATA LOCAL
1 #RESULTI       (I4)
1 #RESULTA       (A20)
1 #RESULTADYN    (A) DYNAMIC
1 #A(I4)         CONST <1234>
1 #B(A20)        CONST <H'30313233'> /* '0123' stored
1 #C(I2/1:3)     CONST <2000, 2100, 2200>
END-DEFINE
*
#RESULTA     := *MAXVAL((IR=A20)         #A, #B)     /*no error, I4->A20 is allowed!
#RESULTADYN  := *MAXVAL((IR=(A)DYNAMIC)  #A, #B)     /*result is (A) dynamic
/* #RESULTI  := *MAXVAL((IR=I4)          #A, #B)     /*compiler error, because conv. A20->I4 is not allowed!  
#RESULTI     := *MAXVAL((IR=I4)          #A, #C(*))  /*maximum of the array is evaluated
DISPLAY #RESULTA #RESULTADYN (AL=10) #RESULTI
END

Resulting Format/Length Conversion Rule Tables

There are different ways to define the resulting format/length of the whole *MINVAL/*MAXVAL system function.

Explicit Specification of the Resulting Format/Length

The resulting format/length of the whole *MINVAL/*MAXVAL system function may be specified by the IR clause. All operands specified will be converted into this resulting format/length, if this is possible without any loss of precision. Afterwards the minimum/maximum of all the converted operands will be evaluated and one single scalar value with the evaluated format/length will be set as result of the whole system function.

Implicit Specification of the Resulting Format/Length

If no IR clause is used inside the *MINVAL/*MAXVAL system function, the resulting format/length will be evaluated regarding the format/length of all operands specified as arguments inside the *MINVAL/*MAXVAL system function. The format/length of each operand is taken and combined with the format/length of the next following operand of the argument list. The resulting format/length of two single operands are then evaluated using the Format/Length Conversion Rule Tables below.

The Format/Length Conversion Rule Table is separated into two different subtables. All combinations not shown in the two tables below are invalid and must not be applied inside the argument list of the *MINVAL/*MAXVAL system function. The keyword FLF indicates that the IR clause must be used in order to define the resulting format/length, because there otherwise may be a loss of precision.

Table 1

Covers all the numeric combinations of two different operands.

    Second Operand
Format- length I1 I2 I4 Pa.b, Na.b F4, F8
First Operand I1 I1 I2 I4 Pmax(3,a).b F8
I2 I2 I2 I4 Pmax(5,a).b F8
I4 I4 I4 I4 Pmax(10,a).b F8
Px.y, Nx.y Pmax(3,x).y Pmax(5,x).y Pmax(10,x).y

if max(x,a) + max(y,b) <= 29
Pmax(x,a).max(y,b)
else FLF

if y=0 and x <=15;
F8
else FLF

F4, F8 F8 F8 F8

if b=0 and a <=15
F8
else FLF

F8

Legend:

FLF Format-length declaration forced. The resulting format must be specified using the IR clause.
Ix Format/length is Integer. x specifies the number of bytes which are used to store the Integer value.
Fx Format/length is Float. x specifies the number of bytes which are used to store the Float value.

Px.y
Pa,b

Packed format with corresponding number of digits before the decimal point (x,a) and the precision (y,b).

Nx.y
Na,b

Numeric format with corresponding number of digits before the decimal point (x,a) and the precision (y,b).
Pmax(c,d).e The resulting format is packed. The length is evaluated by the information following. The number of digits before the decimal point is the maximum value of c and d. The precision value is e.
Pmax(c,d).max(e,f) The resulting format is packed. The length is evaluated by the information following. The number of digits before the decimal point is the maximum value of c and d. The precision value is the maximum value of e and f.

Table 2

Covers all other formats and lengths which may be used for *MINVAL/*MAXVAL system function operands.

  Second Operand
Format-length D T Aa, A dynamic Ba, B dynamic Ua, U dynamic
First Operand D D T NA NA NA
T T T NA NA NA
Ax, A dynamic NA NA A dynamic A dynamic U dynamic
Bx, B dynamic NA NA A dynamic B dynamic U dynamic
Ux, U dynamic NA NA U dynamic U dynamic U dynamic

Legend:

NA This combination is not allowed.
D Date format.
T Time format.
Bx, Ba Binary format with length x, a.
Ax, Aa Alphanumeric format with length x, a.
Ux, Ua Unicode format with length x, a.
B dynamic Binary format with dynamic length.
A dynamic Alphanumeric format with dynamic length.
U dynamic Unicode format with dynamic length.

Evaluating the result-format-length

Using the rules described above, the compiler is able to process the source operands by regarding pairs of operands and calculating an intermediate result for each pair. The first pair consists of the first and the second operand, the second pair of the intermediate result and the third operand, etc. After all operands have been processed, the last result shows the comparison of format and length which will be used to compare all operands in order to evaluate the minimum/maximum. When you use this method of format-length evaluation, the operand format-lengths can appear in any order.

Example: 

DEFINE DATA LOCAL
1 A (I2)      INIT <34>
1 B (P4.2)    INIT <1234.56>
1 C (N4.4)    INIT <12.6789>
1 D (I1)      INIT <100>
1 E (I4/1:3)  INIT <32, 6745, 456>
1 #RES-MIN (P10.7)
1 #RES-MAX (P10.7)
END-DEFINE
*
MOVE *MINVAL(A, B, C, D, E(*)) TO #RES-MIN
MOVE *MAXVAL(A, B, C, D, E(*)) TO #RES-MAX 
DISPLAY #RES-MIN #RES-MAX
END

Output:

         #RES-MIN                      #RES-MAX
---------------------------  ---------------------------

                 12.6789000                 6745.0000000

The following table shows the single steps evaluating the format/length of the example automatically. It shows the intermediate result (ir) of all steps and the comparison format/length (cf) which is used as result-format/length.

Evaluation Order Name of First Operand Format/Length of First Operand or Intermediate Result Name of Second Operand Format/Length of Second Operand or Intermediate Result Format/Length of the Intermediate Result (ir)
1. A I2 B P4.2 ir1 = P5.2
2. ir1 P5.2 C N4.4 ir2 = P5.4
3. ir2 P5.4 D I1 ir3 = P5.4
4. ir3 P5.4 E I4 cf = P10.4

During runtime, all operands are converted into the cf format/length; then all converted values are compared, and the corresponding minimum/maximum is evaluated.

Notes:

  1. If only a single operand is specified, result-format-length will be the format/length of this operand.

  2. If a binary operand with a length in the range 1- 4 is specified as an argument inside the *MINVAL/*MAXVAL system function along with an alphanumeric or Unicode operand, the intermediate result (result-format-length) is evaluated to alphanumeric or Unicode format with dynamic length.

    In this case, the value of the binary operand is considered to be a numeric value, which is converted to the result-format-length according to the data transfer rules (the binary numeric value is converted to unpacked format) before the minimum/maximum is evaluated.

    Example:

    DEFINE DATA LOCAL
1 #B4 (B4)  INIT <1>
1 #A10(A10) INIT <"2">
END-DEFINE
WRITE "=" *MAXVAL(#A10, #B4) (AL=60) /* RESULT FORMAT-LENGTH IS (A)DYNAMIC: "2"
WRITE "=" *MINVAL(#A10, #B4) (AL=60) /* RESULT FORMAT-LENGTH IS (A)DYNAMIC: "1"
END

    Intermediate result-format-length (#A10, #B4) is A dynamic.

    So first #A10 is converted into A dynamic as well as #B4 is converted into A dynamic (considering data transfer rules), before the intermediate result of both operands is evaluated.

Format/Length Evaluation Order

The following graphic represents the order in which format and length are evaluated:

Legend:

ir1, ir2, ir3 Intermediate result 1, 2, 3.
cf Resulting comparison format-length.