hi,
I have a requirement in that i need to process a input file
The problem is, the input file sometimes it is coming in dos mode and some times it is coming in unix mode
The script which i have written will process the file only if it is in unix mode and it is not processing if the file is in dos mode
Is there any method to check whether the input file is in dos mode or in unix mode within the script itself (while processing)
So that i can handle the same in my script

Thanks in advance

Regards,
Selvaraj S

first convert the windows file to unix format using

tr -d '\r' < infile > outfile

or

dos2unix infile > outfile

then process.

Selva,

How exactly is the file coming to the server? Is it through FTP? If so, what mode are you using for transferring? ASCII or BINARY?

Note that if the FTP transfer mode is ASCII, then the files might have ^M characters or square boxes appended at the end of each line, which may result in the failure of your script. Hence, BINARY mode of FTP transfer is always preferred.

HTH,

Regards,

Praveen

hi,
i am not asking for tr -d '\r' < infile > outfile

if you are inside shell script, from the script itself you can check whether a file is having execute, read, write permissions

like that i am asking. From within a shell script i want to check whether the input file is in dos mode or unix mode

My intention is in whatever format (dos or unix) they send the file to unix box my script has to check the type and if it is not matching it has to convert it into unix mode with in the shell script while processing



thanks

#!/bin/ksh

# to type ^M, keeping ctrl pressed, press vm

grep "^M" $1


RETURN_CODE=$?
echo $RETURN_CODE

if [ $RETURN_CODE -eq 0 ]; then
        echo "DOS File"
else
        echo "Unix File"
fi

Hi,
Let me explain the problem


The below few lines are from one file called retst.out
I just copied few of the lines by opening the file in vi mode
you can see the [dos] in the lines below (in vi editor)

Input file in dos mode
------------------------

2008110803000200050000011 10.0000 10.0000
2008111503000200050000011 10.0000 10.0000
2008112203000200050000011 10.0000 10.0000
2008112903000200050000011 10.0000 10.0000

"retst.out" [dos] 52L, 3016C


Same file after converting it into unix mode

2008110803000200050000011 10.0000 10.0000
2008111503000200050000011 10.0000 10.0000
2008112203000200050000011 10.0000 10.0000
2008112903000200050000011 10.0000 10.0000
"retst.out" 52L, 2964C



I will use the below lines for converting my file from dos mode to unix mode

tr -d '\15\32' < retst.out > 1.txt
mv 1.txt retst.out


If you see the file above I am seeing any ctl+M (^M) characters in the file
This conversion I am asking.

How to do the same

thanks

Just for clarification: the term "mode" in UNIX usually refers to the permission bits set (read/write/execute). What you mean would be the line termination style (DOS/Windows: \r\n, UNIX: \n, Apple: \r) or the file encoding.
To convert a file from DOS style to UNIX style, the commands suggested by rakeshou should work just fine.