Multivalue dictionaries

Is the format of the data consistent as described in your example? Always 3 numeric followed by an alpha? 

Or is the number of numeric characters beginning the string random? 

OpenQM has the MATCHFIELDS function to handle this situation - but UniVerse doesn't.

I suspect you will need to put this into a subroutine to successively process each multi-value, and then call that subroutine using the SUBR function.

Cheers,

Brian

I usually create a SB/XA Correlative with the SB/XA Expression Syntax, and look what SB/XA creates for a OE Correlative.
In this case I created a Correlative called FIELD1 with the following Syntax:
(<1>[4,2])

what means take the Value of attribute 1 from Index 4 for 2 characters.
This Creates an Universe  Correlative with the following Values

0001 A              
0002 0              
0003 F1             
0004                
0005               
0006                
0007                
0008 F;1;C4;C2;[]   
0009 L              
0010 5             

When I now call LIST FLAGFILE FIELD1 I get the output as you want.

hope this helps.

Thomas

Thanks Thomas,
Very close.  Unfortunately the the data is not consistent.  That is why I was trying to use the MATCHFILED function of 0N0X.  It's basically a string of numbers, a single alpha, and more numbers including a decimal point in the second set of numbers, sometimes .  I need to display the alpha and numbers after the alpha including any decimal points.  It would be nice to modify your suggestion to replace the C4 with a way to index the location of the alpha in the field.  I tried a few things to no avail.

Tek.

Okay, this makes it a bit more complicated.

Unfortunately for this expression SB/XA doesn't create an OE-Correlative  which is usab le withoult SB/XA installed, but I explain how it worked anyway - maybe you get an idea how it could work. (I'm not good in writing UV-Correlatives my own)

I could manage it by createing an additional helper field.

Lets assume we have a file "FLAGFILE" WITH 2 Items 

Item A -> Attribute 1 = 1234A665

Item B -> Attribute 1 = 123G56

The Helper Field got the SB# Expression: (REPL(OCONV(<1>,"MCA"),"_",<1>))

Which basically replaces the non-numeric part of the whole string with an Underscore.  (assuming there is only one Character)

I called The helper Field "FELD4"

The Results are Item A: 1234_665 ; ItemB: 123_56

The coresponding OE-Correlativ unfortunately looks  a bit weird - so you can not use it without SB/XA installed.

The created OE-Correlative unfortunately looks same weird as the one for the helper field, but in the end it worked...

This should work for you.  If you're hitting a field in another file you can use OCONV(FILENAME;X2;1;1).

I

@RECORD<1,2>

Test

5L

S

You could also try something like this in your I-Type

IFS(EQS(FIELDS(WHOLE,'A',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'B',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'C',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'D',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'E',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'F',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'G',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'H',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'I',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'J',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'K',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'L',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'M',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'N',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'O',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'P',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'Q',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'R',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'S',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'T',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'U',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'V',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'W',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'X',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'Y',2,1),REUSE('')),IFS(EQS(FIELDS(WHOLE,'Z',2,1),REUSE('')),WHOLE,CATS(REUSE('Z'),FIELDS(WHOLE,'Z',2,1))),CATS(REUSE('Y'),FIELDS(WHOLE,'Y',2,1))),CATS(REUSE('X'),FIELDS(WHOLE,'X',2,1))),CATS(REUSE('W'),FIELDS(WHOLE,'W',2,1))),CATS(REUSE('V'),FIELDS(WHOLE,'V',2,1))),CATS(REUSE('U'),FIELDS(WHOLE,'U',2,1))),CATS(REUSE('T'),FIELDS(WHOLE,'T',2,1))),CATS(REUSE('S'),FIELDS(WHOLE,'S',2,1))),CATS(REUSE('R'),FIELDS(WHOLE,'R',2,1))),CATS(REUSE('Q'),FIELDS(WHOLE,'Q',2,1))),CATS(REUSE('P'),FIELDS(WHOLE,'P',2,1))),CATS(REUSE('O'),FIELDS(WHOLE,'O',2,1))),CATS(REUSE('N'),FIELDS(WHOLE,'N',2,1))),CATS(REUSE('M'),FIELDS(WHOLE,'M',2,1))),CATS(REUSE('L'),FIELDS(WHOLE,'L',2,1))),CATS(REUSE('K'),FIELDS(WHOLE,'K',2,1))),CATS(REUSE('J'),FIELDS(WHOLE,'J',2,1))),CATS(REUSE('I'),FIELDS(WHOLE,'I',2,1))),CATS(REUSE('H'),FIELDS(WHOLE,'H',2,1))),CATS(REUSE('G'),FIELDS(WHOLE,'G',2,1))),CATS(REUSE('F'),FIELDS(WHOLE,'F',2,1))),CATS(REUSE('E'),FIELDS(WHOLE,'E',2,1))),CATS(REUSE('D'),FIELDS(WHOLE,'D',2,1))),CATS(REUSE('C'),FIELDS(WHOLE,'C',2,1))),CATS(REUSE('B'),FIELDS(WHOLE,'B',2,1))),CATS(REUSE('A'),FIELDS(WHOLE,'A',2,1)))

I know it's ugly, but it should work.  It's even uglier if you allow upper-case AND lower-case.  The above only assumes upper-case.

WHOLE is a D-Type
00 WHOLE
01 D
02 1
03
04
05 20L
06 M

------------------------------
Tyrel Marak
Technical Support Manager
Aptron Corporation
Florham Park NJ US
------------------------------

As I noted yesterday, the best way to do this is via a subroutine, and the SUBR function. Here is the subroutine:

SUBROUTINE XXTRACT(ans, indata, pattern, elno)
*****************************************************************************
ans = ''
dcv = DCOUNT(indata<1>, @VM)
FOR vno = 1 TO dcv
  adata = indata<1, vno>
  aval = MATCHFIELD(adata, pattern, elno)
  ans<1, vno> = aval
NEXT vno

RETURN
END

And to use this, create a dictionary like this:

I
SUBR('XXTRACT', F1, '0N0X', 2)

XTract
7L
M

This should do exactly what you want.

HTH

Brian

------------------------------
Brian Speirs
Senior Analyst - Information Systems
Rush Flat Ltd
Wellington NZ
------------------------------

There are a number of suggestions below, but here's my option. In the UV home directory create the following entry in APP.PROGS:

     1111
0001 SUBROUTINE U1111(ANS,STATUS,DATA,TYPE)
0002 $OPTIONS DEFAULT
0003 vmcnt = dcount(DATA<1>,@VM)
0004 ANS = ''
0005 for i = 1 to vmcnt
0006    ANS<i> = DATA<1,i>[index(DATA<1,i>,oconv(DATA<1,i>,'MCA')[1,1],1),len(DATA<1,i>)]
0007 next i
0008 STATUS = 0
0009 RETURN
0010 END

Compile and globally catalog this as follows:

BASIC APP.PROGS 1111
CATALOG APP.PROGS $1111 1111

Finally, create the following dictionary for your file:

     A1
0001 I
0002 OCONVS(@RECORD<1>,'U1111')
0003
0004
0005 10L
0006 M

Now, when you use this dictionary you get the result you want (if I understand what you're after):

LIST TESTFILE F1 A1 08:06:47am  13 Dec 2023  PAGE    1
TESTFILE...................... F1............. A1........

0001                           0A10            A10
                               04B9.80         B9.80
                               123Z56568       Z56568

------------------------------
Martin Shields
Senior Technical Consultant
Meier Business Systems PTY LTD
Carnegie VIC AU
------------------------------