I have a master file that connects to a flat file, that's been in use in Version 7.7 for years. I've moved it to Version 8 and it's skipping characters. Here's a sample:

FILENAME=TARIS_IMPORT, SUFFIX=FIX     ,
    DATASET=C:\T_DRIVE\WebFocusFTP\TarisData\TARISDLY.TXT, $
  SEGMENT=TARIS_IMPORT, SEGTYPE=S0, $
    FIELDNAME=RECORD_TYPE, USAGE=A2, ACTUAL=A2, $
    FIELDNAME=TRANS_TYPE, USAGE=A2, ACTUAL=A2, $
    FIELDNAME=TRANS_LOCATION_ID, USAGE=A3, ACTUAL=A3, $
    FIELDNAME=TRANS_OPER_ID, USAGE=A3, ACTUAL=A3, $
    FIELDNAME=TRANS_JULIAN_YY, USAGE=A2, ACTUAL=A2, $
    FIELDNAME=TRANS_JULIAN_DDD, USAGE=A3, ACTUAL=A3, $
    FIELDNAME=TRANS_SEQ_POSITION, USAGE=A3, ACTUAL=A3, $
    FIELDNAME=TRANS_DATE, USAGE=A10, ACTUAL=A10, $
    FIELDNAME=TRANS_TIME, USAGE=A6, ACTUAL=A6, $
    FIELDNAME=EFFECTIVE_DATE, USAGE=A10, ACTUAL=A10, $
    FIELDNAME=WORKSTATION_ID, USAGE=A2, ACTUAL=A2, $
    FIELDNAME=MULTIPLE_IND, USAGE=A1, ACTUAL=A1, $
    FIELDNAME=LIEN_REC_FLAG, USAGE=A1, ACTUAL=A1, $
    FIELDNAME=VOID_CORR_FLAG, USAGE=A1, ACTUAL=A1, $

This is fixed-column data where each field occupies its space on the data line. The first field works correctly, pulling the first two characters. The second field pulls characters 7 and 8 instead of 3 and 4. The fields that follow skip various numbers of characters each time.

When I try to build the mfd from scratch using the synonym builder I get the exact same sort of formatting, and get the exact same error when I run.

Did I change a setting on the server back in 2012 that fixes this and don't remember? Or does Version 8 just do things differently and I'm missing something?

This message has been edited. Last edited by: <Emily McAllister>,

More clues.

I replaced ACTUAL=A2 with ACTUAL=A2B and I got it to read the entire line correctly as it did in Version 7.

This message has been edited. Last edited by: John_Edwards,

John,

I REPROed your example.
Master:

  
FILENAME=JED, SUFFIX=FIX     ,
    DATASET=FOCALPOINT/JED.TXT, $
  SEGMENT=JED, SEGTYPE=S0, $
    FIELDNAME=RECORD_TYPE, USAGE=A2, ACTUAL=A2, $
    FIELDNAME=TRANS_TYPE, USAGE=A2, ACTUAL=A2, $
    FIELDNAME=TRANS_LOCATION_ID, USAGE=A3, ACTUAL=A3, $
    FIELDNAME=TRANS_OPER_ID, USAGE=A3, ACTUAL=A3, $
    FIELDNAME=TRANS_JULIAN_YY, USAGE=A2, ACTUAL=A2, $
    FIELDNAME=TRANS_JULIAN_DDD, USAGE=A3, ACTUAL=A3, $
    FIELDNAME=TRANS_SEQ_POSITION, USAGE=A3, ACTUAL=A3, $
    FIELDNAME=TRANS_DATE, USAGE=A10, ACTUAL=A10, $
    FIELDNAME=TRANS_TIME, USAGE=A6, ACTUAL=A6, $
    FIELDNAME=EFFECTIVE_DATE, USAGE=A10, ACTUAL=A10, $
    FIELDNAME=WORKSTATION_ID, USAGE=A2, ACTUAL=A2, $
    FIELDNAME=MULTIPLE_IND, USAGE=A1, ACTUAL=A1, $
    FIELDNAME=LIEN_REC_FLAG, USAGE=A1, ACTUAL=A1, $
    FIELDNAME=VOID_CORR_FLAG, USAGE=A1, ACTUAL=A1, $

Data:

  
1234567890123456789012345678901234567890123456789
AABBCCCDDDEEFFFGGGHHHHHHHHHHIIIIIIJJJJJJJJJJKKLMN

fex:

  
-* File jed01.fex
TABLE FILE JED
PRINT *
END

Output:

  
  RECORD_TYPE  TRANS_TYPE  TRANS_LOCATION_ID  TRANS_OPER_ID  TRANS_JULIAN_YY  TRANS_JULIAN_DDD  TRANS_SEQ_POSITION  TRANS_DATE  TRANS_TIME  EFFECTIVE_DATE  WORKSTATION_ID  MULTIPLE_IND  LIEN_REC_FLAG  VOID_CORR_FLAG
  -----------  ----------  -----------------  -------------  ---------------  ----------------  ------------------  ----------  ----------  --------------  --------------  ------------  -------------  --------------
  12           34          567                890            12               345               678                 9012345678  901234      5678901234      56              7             8              9             
  AA           BB          CCC                DDD            EE               FFF               GGG                 HHHHHHHHHH  IIIIII      JJJJJJJJJJ      KK              L             M              N             

The output seems fine to me.
Do you have some special characters in your data?

No special characters, and the skipped characters are dependable. That is, my first A2 field claims five characters on every line. When I change to the A2B for the ACTUAL data type the issue clears, A2B claims only 2 characters.

Your result is from version 8? I'm running Version 8 I-Way for this selection.

I'm kind of hoping that this is Version 8 tightening up the spec and this is a dependable solution, because I have two more clients due to go through Version 8 upgrade and "sometimes" isn't a concept I enjoy very much in coding.

John,

I looked in the documentation and I did not find an ACTUAL=AnB. What does it mean?
I am running WF 8.105. Could there be a problem with iWay?

The B presumably stands for Byte and it indicates how many bytes the field consumes. I thought I found it via the synonym editor, but I just went to look to cite it here and I can't find it. I may have found it by searching tech support. I'll tell you this -- I didn't think of it myself.

This may be a setup issue, because the master file I use on the old system was created in I-Way version 7, but runs on I-Way version 8 (release 8008M) on the old system. The new machine is release 81M.

This message has been edited. Last edited by: John_Edwards,

There's more going on here. I have a Dev server running the same source code, in theory an identical install, and I don't need to jump through these hoops. So there's something on this server that's not set up properly. I'm going to open a ticket.

Update -- ticket is open, but it appears to be all about the Code_Page setting in Native Language Support. It was set to Unicode on I-Way (likely for a reason, need to figure out why) but to 137 on WebFOCUS. I believe 137 is the default value.

The old master file now works, and in addition to that I'm getting MUCH faster write speeds to SQL*Server. Apparently the Unicode code_base really scrambles SQL*Server's brain.

Still experimenting with this.

This message has been edited. Last edited by: John_Edwards,