UNIX for Dummies Questions & Answers

yes godlike is right the command shouldn't extract numbers
sorry for not specifying that
can you modify to extract only words please?

It already extracts words - where a word is a sequence of non-whitespace characters separated by whitespace.

You need to exactly define what you mean by 'word'. Just alphabetic characters? (so no words with hyphens or apostrophes, no punctuation, no acronyms with numbers, no actual numbers in the text, etc.)

If you want to strip out everything but alphabetic and whitespace, try:

Code:

sed 's/[^[:alpha:][:space:]]//g' inputfiles | awk 'BEGIN{OFS="\n"} {$0=tolower($0);$1=$1;print}' | sort -u

or

Code:

sed 's/[^[:alpha:][:space:]]//g' inputfiles | tr '[A-Z ]' '[a-z\n]' |sort -u

thank you very much
first commands works but i have 1 last question
if i have /home/user/file the command extract homeuserfile as a word and i need to extract that in 3 words (home user file) can you help please?

You could change the sed command to replace with a space rather than nothing. Or add another sed that just replaces / with space.