Why is sed putting substituted text at the start of a line?

Question

On Mac OS X (bash), I\'m using the following sed at the command line:

$ sed \'s/\\(\\S*\\)/\\1 ~ /\' file1.txt > file2.txt

file1.txt

Answer 1

Pretty sure your sed does not understand \S to mean "non-whitespace". It's most likely trying to match "zero or more S characters", and it finds that match at the beginning of the string.

You want to stick to the POSIX character classes:

sed 's/[^[:blank:]]\+/& ~ /' file1.txt