I want to split a text with comma
for foo in list
for word in $(cat CSV_File | sed -n 1'p' | tr ',' '\n')
do echo $word
Using a subshell substitution to parse the words undoes all the work you are doing to put spaces together.
cat CSV_file | sed -n 1'p' | tr ',' '\n' | while read word; do echo $word done
That also increases parallelism. Using a subshell as in your question forces the entire subshell process to finish before you can start iterating over the answers. Piping to a subshell (as in my answer) lets them work in parallel. This matters only if you have many lines in the file, of course.