IBM Mainframe Forum

by **BillyBoyo** » Fri Aug 16, 2013 3:53 am

A less CPU-strenuous version:

01 WS-VARS.
05 WS-DATA PIC X(44)
VALUE '11245668796981831928381768678756'.
05 WS-REDEF REDEFINES WS-DATA.
10 WS-A PIC 9(09).
10 WS-B PIC 9(07).
10 WS-C PIC 9(07).
10 WS-D PIC 9(07).
10 WS-E PIC X(14).
05 WS-B1-CALC PIC 9(09).
05 FILLER REDEFINES WS-B1-CALC.
10 WS-A1-REMAINDER PIC 99.
10 WS-B1 PIC 9(07).
05 WS-C1-CALC PIC 9(09).
05 FILLER REDEFINES WS-C1-CALC.
10 WS-B1-REMAINDER PIC 99.
10 WS-C1 PIC 9(07).
05 WS-D1-CALC PIC 9(09).
05 FILLER REDEFINES WS-D1-CALC.
10 WS-C1-REMAINDER PIC 99.
10 WS-D1 PIC 9(07).
05 WS-E1-STRING PIC X(16).
05 FILLER REDEFINES WS-E1-STRING.
10 WS-D1-REMAINDER PIC 99.
10 WS-E1 PIC X(14).
05 WS-E1-CALC PIC 9(16).
05 WS-E-NUM PIC 9(14).
05 WS-MODULUS-VALUE PIC 9(02) VALUE 97.
05 WS-RESULT PIC 9(02).
05 WS-E-DIGITS PIC 9(02).
05 WS-DIVISOR PIC 9(09).
05 WS-DIVIDEND PIC 9(09).
PROCEDURE DIVISION.

MOVE WHEN-COMPILED TO W-WHEN-COMPILED
DIVIDE WS-A BY WS-MODULUS-VALUE
GIVING WS-DIVISOR
REMAINDER WS-A1-REMAINDER
MOVE WS-B TO WS-B1
MOVE WS-C TO WS-C1
MOVE WS-D TO WS-D1
MOVE WS-E TO WS-E1
DIVIDE WS-A BY WS-MODULUS-VALUE
GIVING WS-DIVISOR
REMAINDER WS-B1-REMAINDER
DIVIDE WS-B1-CALC BY WS-MODULUS-VALUE
GIVING WS-DIVISOR
REMAINDER WS-B1-REMAINDER
DIVIDE WS-C1-CALC BY WS-MODULUS-VALUE
GIVING WS-DIVISOR
REMAINDER WS-C1-REMAINDER
DIVIDE WS-D1-CALC BY WS-MODULUS-VALUE
GIVING WS-DIVISOR
REMAINDER WS-D1-REMAINDER
DIVIDE WS-D1-CALC BY WS-MODULUS-VALUE
GIVING WS-DIVISOR
REMAINDER WS-D1-REMAINDER
UNSTRING WS-E1-STRING DELIMITED BY SPACE
INTO WS-E1-CALC
DIVIDE WS-E1-CALC BY WS-MODULUS-VALUE
GIVING WS-DIVISOR
REMAINDER WS-RESULT
DISPLAY 'FIRST DIVISION ' WS-A ' ' WS-A1-REMAINDER
DISPLAY 'SECOND DIVISION ' WS-B ' ' WS-B1-REMAINDER
DISPLAY 'THIRD DIVISION ' WS-C ' ' WS-C1-REMAINDER
DISPLAY 'FOURTH DIVISION ' WS-D ' ' WS-D1-REMAINDER
DISPLAY 'SECOND DIVISION ' WS-E ' ' WS-RESULT

GOBACK
.

Output is:

Select all

FIRST DIVISION 112456687 28
SECOND DIVISION 9698183 20
THIRD DIVISION 1928381 86
FOURTH DIVISION 7686787 59
SECOND DIVISION 56 39

The remainder is preprended to a copy of each segment of the input. The final one, with a variable number of digits prepended with the previous remainder, is made into a number using UNSTRING to a numeric field. It will even work when the last 14 bytes are blank, because there will always be a numeric value (the remainder from the previous calculation).

by **rajitha_nair01** » Fri Aug 16, 2013 3:35 pm

Robert Sample wrote:. As long as the 31st through 44th digits are not all zero, the code will process correctly.

and what if a series of 22 charachters is provided as an input.

And is there any way to try the same thing without using intrinsic functions?

by **rajitha_nair01** » Fri Aug 16, 2013 4:06 pm

BillyBoyo wrote:A less CPU-strenuous version:

Select all
01 WS-VARS.
05 WS-DATA PIC X(44)
VALUE '11245668796981831928381768678756'.
05 WS-REDEF REDEFINES WS-DATA.
10 WS-A PIC 9(09).
10 WS-B PIC 9(07).
10 WS-C PIC 9(07).
10 WS-D PIC 9(07).
10 WS-E PIC X(14).
05 WS-B1-CALC PIC 9(09).
05 FILLER REDEFINES WS-B1-CALC.
10 WS-A1-REMAINDER PIC 99.
10 WS-B1 PIC 9(07).
05 WS-C1-CALC PIC 9(09).
05 FILLER REDEFINES WS-C1-CALC.
10 WS-B1-REMAINDER PIC 99.
10 WS-C1 PIC 9(07).
05 WS-D1-CALC PIC 9(09).
05 FILLER REDEFINES WS-D1-CALC.
10 WS-C1-REMAINDER PIC 99.
10 WS-D1 PIC 9(07).
05 WS-E1-STRING PIC X(16).
05 FILLER REDEFINES WS-E1-STRING.
10 WS-D1-REMAINDER PIC 99.
10 WS-E1 PIC X(14).
05 WS-E1-CALC PIC 9(16).
05 WS-E-NUM PIC 9(14).
05 WS-MODULUS-VALUE PIC 9(02) VALUE 97.
05 WS-RESULT PIC 9(02).
05 WS-E-DIGITS PIC 9(02).
05 WS-DIVISOR PIC 9(09).
05 WS-DIVIDEND PIC 9(09).
PROCEDURE DIVISION.

MOVE WHEN-COMPILED TO W-WHEN-COMPILED
DIVIDE WS-A BY WS-MODULUS-VALUE
GIVING WS-DIVISOR
REMAINDER WS-A1-REMAINDER
MOVE WS-B TO WS-B1
MOVE WS-C TO WS-C1
MOVE WS-D TO WS-D1
MOVE WS-E TO WS-E1
DIVIDE WS-A BY WS-MODULUS-VALUE
GIVING WS-DIVISOR
REMAINDER WS-B1-REMAINDER
DIVIDE WS-B1-CALC BY WS-MODULUS-VALUE
GIVING WS-DIVISOR
REMAINDER WS-B1-REMAINDER
DIVIDE WS-C1-CALC BY WS-MODULUS-VALUE
GIVING WS-DIVISOR
REMAINDER WS-C1-REMAINDER
DIVIDE WS-D1-CALC BY WS-MODULUS-VALUE
GIVING WS-DIVISOR
REMAINDER WS-D1-REMAINDER
DIVIDE WS-D1-CALC BY WS-MODULUS-VALUE
GIVING WS-DIVISOR
REMAINDER WS-D1-REMAINDER
UNSTRING WS-E1-STRING DELIMITED BY SPACE
INTO WS-E1-CALC
DIVIDE WS-E1-CALC BY WS-MODULUS-VALUE
GIVING WS-DIVISOR
REMAINDER WS-RESULT
DISPLAY 'FIRST DIVISION ' WS-A ' ' WS-A1-REMAINDER
DISPLAY 'SECOND DIVISION ' WS-B ' ' WS-B1-REMAINDER
DISPLAY 'THIRD DIVISION ' WS-C ' ' WS-C1-REMAINDER
DISPLAY 'FOURTH DIVISION ' WS-D ' ' WS-D1-REMAINDER
DISPLAY 'SECOND DIVISION ' WS-E ' ' WS-RESULT

GOBACK
.

Output is:

Select all
FIRST DIVISION 112456687 28
SECOND DIVISION 9698183 20
THIRD DIVISION 1928381 86
FOURTH DIVISION 7686787 59
SECOND DIVISION 56 39

The remainder is preprended to a copy of each segment of the input. The final one, with a variable number of digits prepended with the previous remainder, is made into a number using UNSTRING to a numeric field. It will even work when the last 14 bytes are blank, because there will always be a numeric value (the remainder from the previous calculation).

just to know:

just to tell you. i mite get numbers that are smaller than 31 too.

and this code dint work for

111023200000435195001213180

just an expample i tried.

correct me if i am wrong in using your code.

by **Akatsukami** » Fri Aug 16, 2013 4:12 pm

rajitha_nair01 wrote:And is there any way to try the same thing without using intrinsic functions?

COBOL intrinsic functions will not attack you the moment you take your eye off of them.

by **rajitha_nair01** » Fri Aug 16, 2013 4:32 pm

Akatsukami wrote:
rajitha_nair01 wrote:And is there any way to try the same thing without using intrinsic functions?

COBOL intrinsic functions will not attack you the moment you take your eye off of them.

thats right indeed!!

but we here are asked to use intrinsic functions only if there is no otjer way out. STANDARDS!!!!...

by **BillyBoyo** » Fri Aug 16, 2013 4:37 pm

Since you set out the 9,7,7,7 yourself, I'm a little surpised that you can have fewer than 30 digits, and freely admit I didn't cater for it.

However, by applying the same technique as I used for the last field to any field which may contain (or be entirely) trailing space, you can easily overcome the mis-description.

Continuing with the two-digit remainder after your number has ended is not a problem, but you can also avoid it by testing the last byte of any possible "space" fields for being space, and knowing that you don't then need to go further.

by **BillyBoyo** » Fri Aug 16, 2013 4:41 pm

Yes, you can avoid using intrinsic FUNCTIONs. Since they are only here for the "last part" of your data, you already have a non-FUNCTION way to deal with it, and there are any number of others.

by **Robert Sample** » Fri Aug 16, 2013 4:58 pm

and what if a series of 22 charachters is provided as an input.

And is there any way to try the same thing without using intrinsic functions?

All right, you need to stop and define the rules for us. Your data contained fields of 9, 7, 7, 7 bytes or 30 bytes total with 14 additional bytes after that. If you can have only 22 bytes then the code could be written to check each 7-byte field to be numeric and if not use NUMVAL on it. You would also need to build in some logic to check for all spaces since 22 bytes means the last 7-byte field would be all spaces. Since you are only giving us pieces of the picture, we will let you figure out the code changes to implement such a strategy -- once you've fully defined the rules.

And why would you want to code up without intrinsic functions? They are a part of COBOL just like the STRING statement and in their place just as useful. It would not be hard (merely tedious) to move the 9-byte and 7-byte fields, one byte at a time using reference modification (or arrays), to a variable that is redefined as numeric -- but why bother when the intrinsic functions provide what is needed?

You have solutions to your original problem -- so now you need to stop trying to change the rules on us and go implement one of the solutions.

by **rajitha_nair01** » Fri Aug 16, 2013 5:01 pm

BillyBoyo wrote:Since you set out the 9,7,7,7 yourself, I'm a little surpised that you can have fewer than 30 digits, and freely admit I didn't cater for it.

its not somthing i thought of. thats an algorithm fr somthing that we are workingg here. its just a way to acheive what is required.

by **rajitha_nair01** » Fri Aug 16, 2013 5:09 pm

thanks all for ur replies. will figure out snthing based on all of your solutions.

IBM Mainframe Forum

modulus operation

Re: modulus operation

Re: modulus operation

Re: modulus operation

Re: modulus operation

Re: modulus operation

Re: modulus operation

Re: modulus operation

Re: modulus operation

Re: modulus operation

Re: modulus operation