Suppressing the INVALIS CHARS in a string

Kamlesh Kamal · Posted: Mon Jul 27, 2009 2:10 pm

Hi ,

I am getting values in my report file as . . AND . . MUT FUND my requirement is to supress these.

Like for . . I should have spaces and for . .MUT FUND I should get only MUT FUND

I have tried checking for NULL VALUES ,LOW AND HIGH VALUES but that didn't work.The hex equivalent for the first byte is 00 and for the second is 1E(which measns a record separator in CHAR)

Can anyone tell me how can I suppress it?The value is as under

PROD TYP CD
.. should be spaces here
.. should be spaces here
..MUT FUND should be MUT FUND here like below
MUT FUND

Robert Sample · Posted: Mon Jul 27, 2009 4:55 pm

Is the data coming in from the input file with the invalid values, or are the invalid values only showing up in the output file? If the data is coming in bad, you need to go to whatever process is generating the file and clean up the data there. If your program is generating the bad data, you need to figure out why and change your code to stop doing so.

Kamlesh Kamal · Posted: Mon Jul 27, 2009 5:10 pm

The data is coming from set of tables and that particular field is populated with that kind of value in the table itself.But my requirement is to populate it without those invalid values in my report that I am generating from those tables

Robert Sample · Posted: Mon Jul 27, 2009 5:55 pm

Are the only two bad characters hex '00' and hex '1E'? If so, use reference modification to scan each byte of the field for one of the bad characters and if it is bad, replace it with a space. If there are other values that are not valid, your IF statement could become long.

Kamlesh Kamal · Posted: Mon Jul 27, 2009 6:27 pm

Yes, there are only two bad characters hex "00" and "1E" but can u plz elaborate on how to use the reference modification for scanning the string..and replacing them?

Robert Sample · Posted: Mon Jul 27, 2009 6:28 pm

There is a manuals link at the top of the page. Use it to find the COBOL Language Reference manual and do a search in it for reference modification.

Marso · Posted: Mon Jul 27, 2009 8:41 pm

or you can use INSPECT:

Terry Heinze · Posted: Tue Jul 28, 2009 6:54 am

If there are a lot of them, use the CONVERTING version of INSPECT. I understand it's faster and it can be done with one statement instead of two.

Kamlesh Kamal · Posted: Tue Jul 28, 2009 12:34 pm

Hi,

Thanks so much for your continued help on this problem.This is really critical for me as I have to deliver this piece of code ASAP.

The first inspect worked fine and the first invalid char of "00" in the form of dot was removed however the second one of "1E" is still there as I am getting a compilation error when I use that inspect.

I am attaching the screenshot for the same.Guess the letter X is not recognized.its taken just as a varible and hence the error.

I just tried giving the inspect as INSPECT EGA-NES-PROD-TYP-CD REPLACINGALL '1E' BY SPACE. It didn't work understandably

enrico-sorichetti · Posted: Tue Jul 28, 2009 12:41 pm

please do not post screenshots...
not everybody can see them
it's a waste of space

but the proper approch would be to understand and fix the program which builds the dataset

if the dataset has a fixed layout it would be easier and faster simply to
rebuild the output records with some MOVE' s

Kamlesh Kamal · Posted: Tue Jul 28, 2009 12:45 pm

The reference modification too won't work as its not fixed how many invalid characters of "1E' would be there in the DS.Like I can have four DOTS followed by the value .example : . . . . MUT FUND

Kamlesh Kamal · Posted: Tue Jul 28, 2009 1:50 pm

Thanks guys..

Got it fixed....using the inspect only ..just had to declare a class in the SPECIAL NAMES under the config section of my program.