RE: Trivia: How many lines of code are allowed in an RPG program? (Canadian Postal Code sample)

["Gary Bowers" <gwbowers@xxxxxxxxxxxxx>]

OK everybody. Time for me to toss in 2 cents more.

 

I've been trying to find the front of this topic, and think that it is from
Frank. Take a close look at his CHECK_56 subroutine, and note that there is
an overlap where A7,A8->(A9 or B0)->B1, B2. So multiple inputs (A9 and B0)
need to get mapped to the same sequential number.  Other solutions that have
the 6th position as a multiple of 10 actually generate too many unique
numbers. The following sample solves that. Pardon the wordiness, it's just
stuff I found on the way, and passed on in case anyone else can use it too.
(I'm not sure I understand WHY the original code operates that way, but I
took it as a functional requirement).

 

 

Try the code below on for size. It's not as elegant as some other procedures
(Lim Hock-Chai)'s looks really nice, and (Dan Bale)'s approach looked fine
with the niggling exception of that last digit. But I think it produces the
correct result set. 

I explain some reasons I picked the structures I did. If you already know
this stuff, just ignore the comments.

 

      * Assign a sequential number to Canadian Postal Code for

      * checking if NEXT sequential number is valid.

      * Equivalent logic as shown by requester's CHECK_56 routine.

      * This does what I THINK it should, but I may have missed something.

      * WARNING/Note: CanadaPostalCode(CPC) position 6 uses base(9) {0..8};

      *

      * Note: THIS IS A SAMPLE ONLY, IT CONTAINING NO ERROR CHECKING!

      * (Valid Letter/number/letter/number/letter/numbers are expected)

      *

      * Actually, I changed from V5 %bif %lookup, doing binary searchs

      * on short arrays(12% slower than using simple and CLEARER %scan).

      * On a little model 800, it returned about 1,000,000 in 6 seconds).

      *

      * Clever readers can transform this into a function for ease of use:

      *      if   ( CanCPCSeq(CurrentCPC)= CanCPCSeq(SavedCPC)+1

      *          or CanCPCSeq(CurrentCPC)= CanCPCSeq(SavedCPC) )

      *        then ...do stuff...

      *        else ...do other stuff...

      *

      * I hope some bright readers will test/verify the program limits,

      * and post this as an RPG function, or turn it into an SQL UDF

      * so that it could be used like the SQL statement below:

      *      Select CITY, CANCPC6, CanCPCSeq(CANCPC6) from Custmast

      *

      * This would be a pretty good example of how to expose/use functions.

      *

      * If you do repost a better version, be sure to take out

      * the *LR on each call, since it would be a performance killer.

      *

      *

      * Background:  Canadian Postal Code format

      * Valid sets of numbers for canadian postal code (LNL NLN)

      * Based on original code requirements:

      *        (LNL NLn )

      *  Where (L.L .L. ) has 26 possible values, A=0 thru Z=25

      *  Where ( N  N . ) has 10 possible values, 0=0 thru 9=9

      *  Where (      n ) has  9 possible values, 0=0 thru 8=8,

      *                      and 9 overlaps, so postal value xxxxx9 gets

      *                      the same sequential number as xxxx10.

      *

      * Sample results: these sequential numbers are returnred for

      *                 various Canada Postal Codes.

      * Call pgm ('A0A0A0') -> sequence(       0)

      * Call pgm ('A0A0A1') -> sequence(       1)

      * Call pgm ('A0A0A8') -> sequence(       8)

      * Call pgm ('A0A0A9') -> sequence(       9)    !!! matches CHECK_56

      * Call pgm ('A0A0B0') -> sequence(       9)    !!! matches CHECK_56

      * Call pgm ('A0A0B1') -> sequence(      10)

      * Call pgm ('A0A0B8') -> sequence(      17)

      * Call pgm ('A0A0B9') -> sequence(      18)

      * Call pgm ('A0A0C0') -> sequence(      18)

      * Call pgm ('A0A0C1') -> sequence(      19)

      * Call pgm ('Z9Z9Z8') -> sequence(15818399)

      * Call pgm ('Z9Z9Z9') -> sequence(15818400)

 

      **********************************************************************

 

     D num10           C                   const('0123456789')

     D ltr26           C
const('ABCDEFGHIJKLMNOPQRSTUVWXYZ')

 

     D colval          S              9  0 DIM(6)

     D rtnCPCseq       S                   like(colval)

 

     C     *entry        plist

     C                   parm                    CPC               6
Canadian Postal Code

 

      *                              (inz column values can be faster)

     C                   eval      colval(6)=1
f(000001)=     1

     C                   eval      colval(5)=%len(num10)-1*colval(6)
f(000010)=     9

     C                   eval      colval(4)=%len(ltr26)*colval(5)
f(000100)=   234

     C                   eval      colval(3)=%len(num10)*colval(4)
f(001000)=  2340

     C                   eval      colval(2)=%len(ltr26)*colval(3)
f(010000)= 60840

     C                   eval      colval(1)=%len(num10)*colval(2)
f(100000)=608400

 

     C                   eval      rtnCPCseq  =

     C
colval(1)*(%scan(%SUBST(CPC:1:1):ltr26)-1)+

     C
colval(2)*(%scan(%SUBST(CPC:2:1):num10)-1)+

     C
colval(3)*(%scan(%SUBST(CPC:3:1):ltr26)-1)+

     C
colval(4)*(%scan(%SUBST(CPC:4:1):num10)-1)+

     C
colval(5)*(%scan(%SUBST(CPC:5:1):ltr26)-1)+

     C
colval(6)*(%scan(%SUBST(CPC:6:1):num10)-1)

 

     C     rtnCPCseq     dsply

     C                   SETON                                        LR