Shell Programming and Scripting

---------- Post updated at 10:00 AM ---------- Previous update was at 09:55 AM ----------

I get an syntax errror at the $1 = $1?

Did not find the error by myself.

Thanks
Robert

You're right if that would be the case

Please post the exact command (you know, spaces and newlines matter) and the exact error message you're getting. Just copy/paste what you see in your terminal.

---------- Post updated at 10:09 AM ---------- Previous update was at 10:05 AM ----------

Well, if that's not the case, a simple loopless sed will be sufficient

Code:

sed -n '
  /\[/{s/\[\([^]]*\)\].*/\1/;N;N;s/\n/;/g;p;}
  ' infile

Hi radoulov and Scrutinizer,

checked now both solutions. And they both lead to the the same result.

When I checked radoulov's solutions, I used "awk '/\[/ {sub(/^\[/, x, $1); sub(/\].*/, x, $1) $1 = $1; print}' RS= FS='\n' OFS=\; test.txt", which resulted in an error. If I copied and pasted directly the code, it worked. Have no idea why...

However, thank you very much for your quick help. It saved me a lot of time!!!!

Best
Robert

Good

Because, as I said, newlines matter. You need either a semicolon or a newline between sub(/\].*/, x, $1) and $1 = $1.

You're welcome!

What do you mean with this? How do I add a newline?

Best
Robert

I mean that this:

Code:

awk '/\[/ { 
  sub(/^\[/, x, $1); sub(/\].*/, x, $1)     <-- You need a newline here!
  $1 = $1; print
  }' RS= FS='\n' OFS=\; infile

and this:

Code:

awk '/\[/ { 
  sub(/^\[/, x, $1); sub(/\].*/, x, $1) $1 = $1; print
  }' RS= FS='\n' OFS=\; infile

... are different.
You need a newline after the second sub command.