Friends,
I admit that I am terrible with regex. I need a script to output all words beginning with "$" from a text file. Here's what I've got so far:
cat file.txt | grep "$" | sed <some-regex>
I don't want the whole line, just the word.
Anyone help? :)
Thanks, Richard
On 13 Jul 16:28, Richard Parsons wrote:
Friends,
I admit that I am terrible with regex. I need a script to output all words beginning with "$" from a text file. Here's what I've got so far:
cat file.txt | grep "$" | sed <some-regex>
How about:
sed -e 's#[^$]*($[^[:space:]]*)#\1 #g; s#(.*)([$][[:alpha:]]*).*$#\1\2#; s# #\n#g;' file.txt
Cheers,
On 13 Jul 16:47, Brett Parker wrote:
On 13 Jul 16:28, Richard Parsons wrote:
Friends,
I admit that I am terrible with regex. I need a script to output all words beginning with "$" from a text file. Here's what I've got so far:
cat file.txt | grep "$" | sed <some-regex>
How about:
sed -e 's#[^$]*\(\$[^[:space:]]*\)#\1 #g; s#\(.*\)\([$][[:alpha:]]*\).*$#\1\2#; s# #\n#g;' file.txt
Doesn't work if there's not a $ on a line, though, it then splits that line up... so, a revision later...
sed -e '/[$]/ { s#[^$]*($[^[:space:]]*)#\1 #g; s#(.*)([$][[:alpha:]]*).*$#\1\2#; s# #\n#g; p; }; d;' file.txt
But, Martijn has a neater solution in the form: grep -E -o '$\w+' file.txt
(Dunno if he's going to post that though :)
On 13/07/10 16:54, Brett Parker wrote:
But, Martijn has a neater solution in the form: grep -E -o '$\w+' file.txt
Just beaten to the post! But I'll post my version in case it's useful:
grep -oE '$\w+' file.txt | sort -u
-E enables regex matching -o tells grep to only output the match (not the whole line) sort sorts the results, -u removes duplicates
You can test your regular expressions here: http://gskinner.com/RegExr/ (Flash based, sorry)
Tim.
Thanks everyone so much for your swift responses. Perhaps one day I will be a regex guru too...
Richard
On 13/07/10 16:28, Richard Parsons wrote:
Friends,
I admit that I am terrible with regex. I need a script to output all words beginning with "$" from a text file. Here's what I've got so far:
cat file.txt | grep "$" | sed <some-regex>
I don't want the whole line, just the word.
You can even do it with just grep:
mak@yoda:~$ (echo hi; echo 'now $food fight $win'; echo '$$$'; echo there) | grep -E -o '$\w+' $food $win
Cheers,
-- Martijn
-
Brett Parker -
Mark Rogers -
Martijn Koster -
Richard Parsons -
Tim Green