Attempting to remove string between two strings, whilst omitting the string if - sed

Question

I'm having some trouble when attempting to remove a string between the string end and the character ;. At the moment it looks like I'm also picking up the string if on other lines as that string is also between end and ;. I've been attempting to omit the if string from the command but I don't seem to be having much look using the below sed command as I receive a error. The string triggername is what needs to be removed, but it's not consistently named that.

Example File Before:

end if;
    end triggername ;

Example File Expected:

end if;
    end;

The command I've been attempting to use is:

sed -i -e '/if/! s/\(end.*)\;/g' test

I was hoping the /if/! would omit all literal if strings from the next command but when I attempt to run it I receive the following error:

sed: -e expression #1, char 20: unterminated `s' command

Can anyone help me understand why I'm getting that error please? And if the sed command is mainly correct, if not what would you recommend to improve it.

Update1:

Thank you! As suggested I've tried adding in an extra / at the end as I just require it removing not replaced. But now I have this error: sed -i -e '/if/! s/\(end.*)\;//' test sed: -e expression #1, char 20: Unmatched ( or \(

To me it looks like I have a closing bracket not too sure about the '\' . I have only done this a once before, so quite new.

Answer 1

1. Your first attempt misses the replacement part from s/pattern/replacement/modifiers syntax.

2. You forgot to escape the closing bracket: Use \).

3. No need to escape the semicolon though.

4. Your pattern includes all of end.*;. So all this will get removed. Use \1; as replacement.

In total:

sed -i -e '/if/!s/\(end\).*;/\1;/' file

Or as end is a static word and .* includes the ; you can simplify it:

sed -i -e '/if/!s/end.*/end;/' file