MVSFORUMS.com

[varundb]

Hi

I am getting a file from Unix into a mainframe dataset using FTP. Each record is getting a '.' at the end. When I open up the file in the hex mode, it shows 0D. Can someone please help me strip this character using the FTP utility itself.

Thanks in advance.

[kolusu]

varundb,

Did you get the file to unix from a windows machine? if so that explains as to why you had the extra '0d' at the end. In any case You can delete the x'0D' from unix using the tr command like

[varundb]

Kolusu

Thanks for the quick reply. The problem I am facing is that I don't have admin rights on the Unix box. We get an encrypted file from a third party and the script on Unix box decrypts it from where I get it using FTP. Ideal solution would be if I can do some manipulation in the FTP job run at the mainframe to truncate these characters.

Thanks for your help

[superk]

I'd be curious to know what the files actually contain for both the end-of-line (EOL) and eond-of-file (EOF) values.

Anyway, I'm sure you could pass the file through a SORT after you receive it and truncate the last byte.

[varundb]

Superk

Thanks for the reply, but the problem is that the file I am talking about contains variable length records.

[varundb]

If there is no way of doing that using the FTP utility, can I remove the last character from a file with variable length records using DFSORT or some other utility?

Thanks for all your help..

[varundb]

As an example-
The file looks like

[Frank Yaeger]

Here's a DFSORT job that will do what you asked for:

[varundb]

Frank

Thanks a ton for the quick reply, but I guess my last reply was a little ambiguous.. my file is a fixed length file which can have variable length records, the above step won't work on that.. right?

Regards

[varundb]

Right now I am using this card-
//SYSIN DD *
SORT FIELDS=COPY
OUTREC FIELDS=(01,80,TRAN=ALTSEQ)
ALTSEQ CODE=(0D40)
/*

This is working as it is replacing the Cr with spaces. Is it the best possible solution?

[Terry_Heinze]

"...my file is a fixed length file which can have variable length records,..."
"Fixed length file" means (to me anyway) your file has the same number of records each time and is a fairly meaningless term. VB (variable, blocked) means a file consisting of variable length records and is blocked at more than 1 record per block. See manuals.
_________________
....Terry

[kolusu]

varundb,

You can use ALTSEQ code if you don't have negative packed decimal number. Negative packed decimal numbers have x'd' in the last nibble and you will be replacing it with a space.

Terry,

OP is having a fixed block file but the position of CR is variable. like shown below

[Terry_Heinze]

"OP is having a fixed block file but the position of CR is variable."
This sentence seems self-contradictory to me. FB means fixed length records; VB means variable length records. OP: Could you verify that your records are of variable length? And not use the term (F)ixed (B)locked file? See definitions of FB and VB in the JCL Manual.
_________________
....Terry

[varundb]

Terry

Sorry for all the confusion. When I say the file is fixed length, that means it is defined as a FB file with length(let's say 80). But the records can be anywhere from 50 to 80 characters long(there is a delimiter in the file). So if for the first record, the amt is 10$, the record will be something like 01|123|10. and if for the second record, the amount is 100,000, then the record would be 02|122|100000.

The file length is defined to be the maximum length that can be.

As for Kolusu's question, the file will have the amounts as 9(3).9(2)(It will always be positive).

Thank you all for your help in this.

[Frank Yaeger]