[[https://www.revelation.com/|Sign up on the Revelation Software website to have access to the most current content, and to be able to ask questions and get answers from the Revelation community]]


==== Unique combinations of numbers in a set (AREV Specific) ====
=== At 06 AUG 2011 06:44:07AM William Titus wrote: ===
 
{{tag>"AREV Specific"}}
 
<QUOTE>
From a sorted list of numbers such as "10,20,30,40,50" I need to build an ordered list of all unique combinations without repeating elements.



In this case, the combinations would be:



10, 20, 30, 40, 50, 10*20, 10*30, 10*40, 10*50, 20*30, 20*40, 20*50, 30*40, 30*50, 40*50, 10*20*30, 10*20*40, 10*20*50, 10*30*40, 10*30*50, 10*40*50, 10*20*30*40, 10*20*30*50, 10*30*40*50, 10*20*30*40*50



The number of numbers is not fixed.



Can one of you math wizards set me on the right programmatic path (recursive or otherwise) to get me where I need to go?



Thanks!
</QUOTE>
----

=== At 06 AUG 2011 11:27AM Richard Hunt wrote:  ===

<QUOTE>Since I am not sure if you can delimit by using value marks and field marks rather than the commas ans asterisks, I used FOR/NEXT rather than REMOVE statements...



RESULT='

DELIMITER=*'

STRING=1:DELIMITER:2:DELIMITER:3:DELIMITER:4:DELIMITER:5

STRING=1:DELIMITER:2:DELIMITER:3:DELIMITER:4:DELIMITER:5

LAST_VALUE=COUNT(STRING,DELIMITER) + (STRING NE '')

FOR VALUE=1 TO LAST_VALUE

  RESULT := @VM:FIELD(STRING,DELIMITER,VALUE)

NEXT VALUE

FOR VALUE=1 TO LAST_VALUE

  PREFIX=FIELD(STRING,DELIMITER,1,VALUE)

  SUFFIX=FIELD(STRING,DELIMITER,VALUE + 1,LAST_VALUE)

  LAST_SUFFIX_VALUE=COUNT(SUFFIX,DELIMITER) + (SUFFIX NE '')

  FOR SUFFIX_VALUE=1 TO LAST_SUFFIX_VALUE

    RESULT := @VM:PREFIX:DELIMITER:FIELD(SUFFIX,DELIMITER,SUFFIX_VALUE)

  NEXT SUFFIX_VALUE

NEXT VALUE

RESULT=RESULT2,LEN(RESULT)

</QUOTE>

----

=== At 06 AUG 2011 02:58PM William Titus wrote:  ===

<QUOTE>Thanks, Richard, for taking time to send me your code. It's a lot more compact than mine and took me as far as:



1²2²3²4²5²1*2²1*3²1*4²1*5²1*2*3²1*2*4²1*2*5²1*2*3*4²1*2*3*5²1*2*3*4*5



The code I've built to this point is this:



***********



DATUM="

RUN=1

NUMBERS=1,2,3,4,5"

CONVERT "," TO @VM IN NUMBERS

HOWMANY=COUNT(NUMBERS,@VM)+(NUMBERS NE "")

COMBOS=HOWMANY * HOWMANY

ARRAY="

FOR X=1 TO HOWMANY

  FOR G=1 TO HOWMANY

  ARRAY=NUMBERS

  NEXT G

NEXT X

*

ALLCOMBOS="

FOR X=1 TO COMBOS

 FOR G=1 TO HOWMANY

   DATUM=ARRAY

   LOCATE DATUM IN ALLCOMBOS USING @VM SETTING POS ELSE

     IF DATUM NE "" THEN

       IF ALLCOMBOS=" THEN ALLCOMBOS=DATUM ELSE ALLCOMBOS:=@VM:DATUM

     END

   END

 NEXT G

NEXT X

*

FOR X=1 TO COMBOS

  D=NUMBERS

  FOR Y=X+1 TO COMBOS

    E=NUMBERS

    IF E NE "" THEN

      DATUM=D:"*":E

    LOCATE DATUM IN ALLCOMBOS USING @VM SETTING POS ELSE

      IF DATUM NE "" THEN

        IF ALLCOMBOS=" THEN ALLCOMBOS=DATUM ELSE ALLCOMBOS:=@VM:DATUM

      END

    END

    END

  NEXT Y

NEXT X



FOR X=1 TO COMBOS

  D=NUMBERS

  FOR Y=X+1 TO COMBOS

    E=NUMBERS

    FOR Z=Y+1 TO COMBOS

      F=NUMBERS

      IF E NE "" AND F NE "" THEN

        DATUM=D:"*":E:"*":F

      LOCATE DATUM IN ALLCOMBOS USING @VM SETTING POS ELSE

        IF DATUM NE "" THEN

          IF ALLCOMBOS=" THEN ALLCOMBOS=DATUM ELSE ALLCOMBOS:=@VM:DATUM

        END

      END

    END

    NEXT Z

  NEXT Y

NEXT X



***********



And it gets me to:



1²2²3²4²5²1*2²1*3²1*4²1*5²2*3²2*4²2*5²3*4²3*5²4*5²1*2*3²1*2*4²1*2*5²1*3*4²1*3*5²1*4

*5²2*3*4²2*3*5²2*4*5²3*4*5



I realized that this is a kludge and only gets me to 3-digit groups, though it gets exactly what I want to that point.



I think I need a simple recursive way to minimize the code and maximize efficiency. I'm just not proficient enough as a coder at this point to know the way to get there except by adding more FOR-NEXT loops while also accounting for the maximum number of numbers the user will want to put into as many groups as possible.



Thanks again for your help.



</QUOTE>

----

=== At 08 AUG 2011 10:13AM Michael Slack wrote:  ===

<QUOTE>Just off the top of my head, I think you might find a FOR loop and a LOCATE BY to be simpler.  I'll sketch out what I'm thinking in pseudo code.



* Start_Array is your array or row that you want to work with.

* End_Array is your resulting array or row.



CNT_START_ARRAY=COUNT(START_ARRAY, @VM) + (START_ARRAY # '')



IF CNT_START_ARRAY=0 THEN STOP PROCESS



FOR I=1 TO CNT_START_ARRAY



  LOCATE START_ARRAY IN END_ARRAY BY "AL" USING @VM SETTING POS ELSE

    END_ARRAY=START_ARRAY

  END



NEXT I



What this does is any time it doesn't find a matching value in the END_ARRAY, it put it in, in sorted order.  So, if it find a duplicate value already within the END_ARRAY, it won't add it.  So, in the end you get only unique values within the END_ARRAY.



Of course if this is what you're looking for, you'll need to flesh it out to fit your particular needs.



I hope this helps.

Michael Slack

</QUOTE>

----

=== At 10 AUG 2011 02:37PM Mike Ruane wrote:  ===

<QUOTE>Mike-



I'd use the same code as you, except using a LOOP...REMOVE to remove the items from the array, instead of the FOR...NEXT loop.



Mike

</QUOTE>

----

=== At 11 AUG 2011 04:49PM Bill Titus wrote:  ===

<QUOTE>Thanks, all.



Your help is much appreciated by this old dog. Now it's off to implement what's been suggested!

</QUOTE>


[[https://www.revelation.com/revweb/oecgi4p.php/O4W_HANDOFF?DESTN=O4W_RUN_FORM&INQID=NONWORKS_READ&SUMMARY=1&KEY=1407169525032DFC852578E4003AF8D7|View this thread on the  forum...]]