Csv file blank data

prasadplease · New User Joined: 02 Sep 2006 Posts: 31 Location: Mumbai

Hi,
My job emails a sequential file(having ; delimiters) via email into a csv file.
The problem is if any field in any particular record of the file has junk data, all data following that field appears as blank in the csv file.
I can see the data in the sequential file but not in the csv file.

Has anyone faced this problem before?

Thanks,
Prasad.

prasadplease · New User Joined: 02 Sep 2006 Posts: 31 Location: Mumbai

It would also be of help if someone can guide me how to check in SORT if an alphanumeric field has junk or null data

expat · Posted: Wed Aug 12, 2009 4:28 pm

What is the definition of junk ?

Robert Sample · Posted: Wed Aug 12, 2009 4:47 pm

I echo expat: what are you defining as junk characters? I've transferred many dozens of .csv files using tabs, commas, and semicolons as delimiters and never had a problem. But then, I do insist on having quote marks around any alphanumeric fields if there's any chance of non-display data in them.

prasadplease · New User Joined: 02 Sep 2006 Posts: 31 Location: Mumbai

By Junk data i mean non-displayable data (the kind u need to F P'.')

@Robert
I tried using quotes around the particular field, but still facing the problem.

Robert Sample · Posted: Wed Aug 12, 2009 6:13 pm

Actually, you have two problems then. The first is that if the data is not displayable, your transfer of the file to another machine in text mode will do unpredictable things to your data. This is most likely the cause of your original issue. If you transfer the file in binary, it will not be readable on the other machine since .csv files tend to be used on Windows machines.

Therefore, you have two options:
1) convert the non-display data to displayable EBCDIC data on the mainframe before transferring to the other machine -- which will probably resolve your other issue
2) give up on the .csv idea and transfer the file as binary, then use a routine on the other machine to convert -- byte by byte -- the file from EBCDIC to a collating sequence recognized by the other machine.

dick scherrer · Posted: Wed Aug 12, 2009 9:56 pm

Hello,

prasadplease · New User Joined: 02 Sep 2006 Posts: 31 Location: Mumbai

Thanks everyone.

I replaced the junk/non-displayable data from the file using a sort step and then emailed the file as csv.

//SYSIN DD *
SORT FIELDS=COPY
ALTSEQ CODE=(0040)
OUTREC FIELDS=(1,525,TRAN=ALTSEQ)
/*

dick scherrer · Posted: Thu Aug 13, 2009 7:43 pm

Hello,

prasadplease · New User Joined: 02 Sep 2006 Posts: 31 Location: Mumbai

I agree with you Dick that there shud not be bad data in the first place.
Bad dat was in test but I dont expect such data in production. This is just a precautionary step.

dick scherrer · Posted: Thu Aug 13, 2009 9:46 pm

Thanks for the clarification - good luck

d