UNIX for Dummies Questions & Answers

Hi all,

I am trying to write a ksh script to remove controlM characters from the input csv file.

I tried following command

Code:

sed 's/^M//g' $DIR/$FILENAME > $DIR/$Temp_FILENAME

while writing the code i had to press CTRL key then V then M (to type ^M)

Cant I write ^M without pressing ctrl+V ctrl+M?? Problem is when I migrate the code to some other server I have to bring the code to Windows and then to the next Unix server and the ^M char gets lost in windows.

Is there a way that sed command can read ^M without pressing ctrl+ V M manually??. e.g [CTRL]V M
so that i dont have to manually type in ^M in the script in production

Please help.

Thanks

Code:

tr -d \\015 < file

Here is an example that should process all of your .csv files.

Code:

for CSV_FILE in *.csv; 
  do if [ -f $CSV_FILE ]; 
      then sed -e 's/\x0d//g' $CSF_FILE > /tmp/foo.$$; 
      mv /tmp/foo.$$ $CSF_FILE; 
  fi;
done

thanks everyone

following also works fine in the script

Code:

sed 's/'"$(printf '\015')"'$//g' $DIR/$FILENAME > $DIR/$Temp_FILENAME

Thanks
again

To replace the ^M 's on the screen in vi type

:%s/\r//

What that says is: for lines 1 to end, replace \r with nothing . This will replace only one instance of ^M per line. If you have two on a line for some reason simply run the command a second time.

The th^Me thus providing t^Mhe ability to change characters, ^mwords, and lines

I tried doing it in command mode with this text in my vi editor... It doesnt delete ^M, but it threw me an error

E486: pattern not found: r//

Please explain in detail.

Try this

Code:

:1,$ s/[ctrl-key+v and ctrl-key+M]//g 
but actual command would be 
:1,$ s/^V^M//g