Question

I have a file that looks like this:

$variable['keyword1']   = 'value1';
$variable['keyword2']   = 'value2';
$variable['keyword3']   = 'value3';
$variable['keyword4']   = 'value4';

keyword 1-4 are all unique and just letters and numbers.

value 1-4 may or may not be unique and can be any combination of letters, numbers, special characters

I need to be able to extract value1 or value2 or value3 or value4 and save it as $somevariable. So the keyword is obviously the thing that I need to define in the sed command so that I get the desired value out. How would I do that using an sed one liner? Doesn’t look that hard to me but I can’t seem to crack it. Been googling and trying different combinations for hours.

Answer 1

sed -n "s/\$\w\+\['keyword1'\]\s*=\s*'\([^']\+\)';/\1/p" inputfile

to save it:

somevariable=$(sed -n "s/\$\w\+\['keyword1'\]\s*=\s*'\([^']\+\)';/\1/p" inputfile)

if your file is exactly like you describe, you can also use this:

grep keyword1 inputfile | sed "s/.*=\s*'\([^']\+\)';/\1/"

or better yet: (thanks William)

sed "/keyword1/s/.*=\s*'\([^']\+\)';/\1/" inputfile