Hi,

I would like to remove all characters between [ and ] in each line of a text file but only if there is NOT an X between the [ and ]. If there IS an X then the string between the [ and ] should remain unchanged.

Is this possible with sed?

Many Thanks,
Jonny

grep -v "X" FILENAME | sed "s/[0-9,A-Z]//g"

# echo "This [.] is a test" | sed -e '/\[[^X]\]/s/\(\[\)[^]]*\(\]\)/\1\2/'

# echo "This [X] is a test" | sed -e '/\[[^X]\]/s/\(\[\)[^]]*\(\]\)/\1\2/'

David:

Would you mind providing an explanation of your sed solution? Thanks!

Nails

I think the solution needs to allow for any number of characters inside the brackets - not just a single character. And also allow for multiple bracketed expressions in each line:

echo "This [abc][rXst][def] is a test" | sed -e 's/\[[^X[]*\]/[]/g'

I was really hoping someone more clever than me would answer this but I just couldn't leave grep -v as the answer.

sed -e '/\[[^X]\]/s/\(\[\)[^]]*\(\]\)/\1\2/

/\[^X]\]/ match a line that does not have X between the [ and ]. This should be done better.

/\(\[\)[^]]*\(\]\)/ grab the [ into \1 and ] into \2 omitting the string of anything other than ] in between.

My solution posted earlier searches for a string that begins with [ that is followed by any number of characters that are NOT X or [ and terminates with ]. And it will change that string to []. The /g flag makes it process all occurrences of this pattern on each line instead of just the first occurrence.

I exclude the [ in my pattern because an RE will always qualify the longest possible string. If I did not exclude the [, then it would locate a single pattern in the string:

abc[def]ghi[jkl]mno

and change it to:

abc[]mno

instead of the correct:

abc[]ghi[]mno

Many thanks to all of you for replying.

Jim's one seems to be working.

Regards,
Jonny

If anyone knows how to append a blank line after every
single space - I would greatly appreciate it.

ie:

before:

hello world how are you?

after:

hello
world
how
are
you?

Thank you in advance!  

echo "hello world how are you?"|xargs -n1

thank you for the response - but i should have been
clearer. i have a file with a list of items separated by
spaces - and i want to have each item on a separate line.

i tried

cat filename|xargs -nl > filename2

and i get

xargs: illegal argument count

i am using the terminal on OS X - is that a factor?

excuse my ignorance - and thanks again!

fxcre8or