UNIX for Dummies Questions & Answers

I have a file in which I need to remove all newline characters but only if they are preceded by "^C" ( ASCII value 3 )

Basically this is a bulk copy file with "^B" and "^C" as delimiters for each record. The application producing the file is also appending a newline which I need to remove, however there is the potential for newline characters to also exist within the body of each record and these must not be touched.

E.G. If myfile contains :

^Babc123^C
^Bxyz
987^C
^Bfoobar^C

I need to convert to :

^Babc123^C^Bxyz
987^C^Bfoobar^C

See how it maintains the newline after "xyz" but supresses the others ?

Obviously I can't use tr '\010' '' because that gets rid of the embedded linefeed that I have to keep. I tried to use sed as follows :

cat myfile | sed 's/^C\\n/^C/' > mynewfile

That doesn't work because apparently sed strips the newline for each record it reads in before applying the substitution.

Any bright ideas ? I'm guessing that it may need some black art like nawk

Try this - note not all seds allow \xnnn for a hex number:

Code:

sed -e :a -e '/\x003$/N; s/\x003\n//; ta' oldfile > newfile