Handle junk values not readable by COBOL

surya anem · New User Joined: 20 Dec 2007 Posts: 54 Location: Hyderabad

I have a file that is transferred from a non-mainframe system to a mainframe system. When the characters in ASCII are transitioned to EBCDIC format, spaces (HEX '40') are getting converted to junk values (HEX '05') which is not readable by COBOL.

Is there any way to handle this in cobol.

Please let me know if you need more info.

Craq Giegerich · Posted: Tue Jul 13, 2010 5:21 pm

surya anem · New User Joined: 20 Dec 2007 Posts: 54 Location: Hyderabad

Please see below the field that is being said as invalid when it is viewed through file-aid. If you can detect any, it would be of great help.

dbzTHEdinosauer · Posted: Tue Jul 13, 2010 5:36 pm

surya anem,

since we have no idea of the field definitions, what you posted above is useless.

surya anem · New User Joined: 20 Dec 2007 Posts: 54 Location: Hyderabad

This is just one field
WS-NAME PIC X(108)

I just posted the beginning portion of the field that has data. rest of them were spaces (HEX '40')

Craq Giegerich · Posted: Tue Jul 13, 2010 5:41 pm

I don't see a x'05' anywhere in that data, all I see are EBCDIC characters, mostly spaces x'40'. What is the definition of the field, which field is INVALID?

dbzTHEdinosauer · Posted: Tue Jul 13, 2010 5:42 pm

well, a pic x can not contain invalid characters.

you need to explain your problem.

saying you have junk characters only tells us that you have no idea what is going on.

the must be a file definition.
what about the file does not follow the definition?

surya anem · New User Joined: 20 Dec 2007 Posts: 54 Location: Hyderabad

The data i posted is value of field WS-NAME PIC X(108) in 2 records (1 above line and 1 below line) when viewed in file-aid with HEX ON.

Please see the hex value for the space between Novello and Dan.

I hope i am clear now. if still not, plz let me know.

dbzTHEdinosauer · Posted: Tue Jul 13, 2010 5:51 pm

use inspect replacing in your cobol program,
or pre-process the file with a sort
to change the 05 to 40, a space.

what is the problem?

if there is an FTP problem, you need to check with your support people.

and it is possible, that there was an ASCII tab character in the original file prior to the transfer?
sorta depends on how the ASCII file was generated.

Robert Sample · Posted: Tue Jul 13, 2010 5:56 pm

surya anem · New User Joined: 20 Dec 2007 Posts: 54 Location: Hyderabad

We recieve this file from non-mainframe through FTP.

the problem is when i view this file through file aid (with hex off) only this field is being displayed as invalid rest all look fine.

As you pointed, i am not exactly sure why this is happening. I just want to know if there is any specific reason for this and any ways i can handle this to prevent this invalid values (which i guess are non-ebcdic characters that got into file during FTP) from being passed on to next step.

So i just want to filter this out and populate only valid values.

I suspected this HEX '05' because rest of the spaces are coming as HEX '40'

Craq Giegerich · Posted: Tue Jul 13, 2010 6:06 pm

I am sorry there is a x'05' character in there but cobol can handle that with no trouble. What program is saying that the field is invalid, what is causing it to say it is invalid? Please tell us something other then you don't know what you are doing.

Robert Sample · Posted: Tue Jul 13, 2010 6:07 pm

The easiest answer: DO NOT USE FILE AID TO VIEW THE FILE.

You have not indicated any problem with any program actually using this data -- just viewing it via File Aid. So don't view it and end of issue.

dbzTHEdinosauer · Posted: Tue Jul 13, 2010 6:14 pm

I doubt that file-aid would call the contents of an pic x field invalid,
unless there is a way to tell file-aid to only consider 'printable char'.

do the records of this file only contain 1 field?
what other fields make up the record.

Besides, it appears that the x'05' would be a handy delimiter to separate the last name from the first in an UNSTRING operation.

surya anem · New User Joined: 20 Dec 2007 Posts: 54 Location: Hyderabad

Thats a good solution. But i was just trying to understand why and what values does file-aid considers invalid in an ALPHANUMERIC field for viewing.

CICS Guy · Posted: Tue Jul 13, 2010 6:23 pm

dbzTHEdinosauer · Posted: Tue Jul 13, 2010 6:48 pm

Robert Sample · Posted: Tue Jul 13, 2010 7:03 pm

In my experience, File Aid only identifies invalid data when you have a NUMERIC (not alphanumeric) variable in a copy book where the data does not match the specification. Of course, if the TAB character is supposed to represent a given number (4? 5? 8?) of spaces, that could cause the rest of the record to not match the copy book.

Bill O'Boyle · Posted: Tue Jul 13, 2010 8:16 pm

In the following link, review the building of the FROM and TO translate tables, followed by an INSPECT and see if this will work for you.

ibmmainframes.com/viewtopic.php?p=143786&highlight=#143786

Bill

dick scherrer · Posted: Tue Jul 13, 2010 8:16 pm

Hello,

If you are trying to view the data in file-aid using a file layout (copybook), suggest you look at the file without this.

The process that created the file placed a tab delimiter between the "lastname" the the first part of the name. I suspect you may need to handle this rather than simply changing the tab to a space - depending on the "rules".

Kjeld · Posted: Thu Jul 15, 2010 1:28 pm

haiardhan · New User Joined: 26 Jul 2008 Posts: 7 Location: india

Hi guys i understood the problem what question is :

x'05' can not shown any value ... not even spaces(x'40')

x'05' is happend to be appear becuase of typo mistake of user.... when he enter the data ... by accidently he might have entered alt+'some key'

hense the data is showing like as he posted.

SWL.I.S..TRFT.SRM.15
EED0C4E40EDCE0EDD0FF
26359B2B539635294515[/b]

Robert Sample · Posted: Wed Apr 24, 2013 5:54 pm

haiardhan, nice of you to join the discussion two years and nine months late ....

haiardhan · New User Joined: 26 Jul 2008 Posts: 7 Location: india

Robert....

I hope surya was giving you a data which is having problem...with x'05' in between the two words of a record.

The data is something like below

OWEN.SLIDE SANDAL
DECD0EDCCC4ECDCCD4
665552394502154130

The junk here is the dot between owen and slide 'OWEN.SLIDE'
which is a junk value of equivalent of hex '05'

please check the below attached for the exact problem to understand.

[/img][/url]

Robert Sample · Posted: Wed Jun 12, 2013 5:17 pm

haiardhan: a few points you need to consider:

1. surya's last post was July 2011. While surya may still be lurking, it is also entirely possible you are providing advice to someone who stopped reading posts 21 months before you posted.

2. there is no such thing as a "junk value" -- period. Each and every character in the collating sequence has a purpose, if nothing else than to complete the collating sequence. Talking about "junk values" indicates that you have ABSOLUTELY no knowledge of computers, OR that your thinking is so sloppy that anything you post is suspect, automatically.

3. as I posted before, the X'05' character represents a tab character. It can be used in TSO/ISPF edit sessions as well as come to a mainframe from a PC in a text file.

4. do not post attachments -- many on this forum cannot download them due to company restrictions, so you are significantly reducing your audience when you post an attachment.

5. what are you trying to say, anyway? If you want to say that X'05' is not a valid character on a 3270 screen, so what? It can still be placed on a screen and it is still not a "junk value".