CODEWITHSUNDEEP
shellawk
Jimmy
Jimmy

Reputation: 881

Return string between single quote with awk

I'm currently writing a script to mass reset wordpress username and passwords. The passwords were stored using old-passwords and now with an update to MySQL and PHP, the old passwords are no longer working and hence need to be reset.

I have a file that looks like this:

define('DB_USER', 'abc_12345');
define('DB_PASSWORD', 'abc12345');
define('DB_USER', 'def_34589');
define('DB_PASSWORD', 'def34589');

I'm just needing to return the values inside the single quotes:

DB_USER abc_12345
DB_PASSWORD abc12345

DB_USER def_34589
DB_PASSWORD def34589

Now, I've looked around at the similar questions, but they all are a bit different then what I'm working with.

I've tried the following:

cat file.txt | awk '{print $1,$2}'

which returns 

define('DB_USER', 'abc_12345');
define('DB_PASSWORD', 'abc12345');
define('DB_USER', 'def_34589');
define('DB_PASSWORD', 'def34589');

I've tried

cat file.txt | awk '{print $1}'

which returns

define('DB_USER',
define('DB_PASSWORD',

I've tried

cat file.txt | awk '{print $2}'

which returns

'abc_12345');
'abc12345');

I've also tried awk mixed with grep

cat file.txt | awk '{print $2}' | grep -P " '*.' "

which doesn't return anything.

Upvotes: 4

Views: 7965

Answers (4)

warybyte
warybyte

Reputation: 196

I was able to accomplish this in Cygwin the following way...

$ cat test.file | awk -F "\'" '{print $2" "$4}'
DB_USER abc_12345

You might get an awk warning because of the escapes, however the output should be exactly what you were wanting using similar syntax to what you already tried. I found that using double-quotes, at least in my shell, around a single-quote delimiter works well as long as you escape the single in the middle. For the output field of awk however you want to switch back to the single-quotes, otherwise you will have errors like this...

$ cat test.file | awk -F "\'" "{print $2" "$4}"
awk: cmd. line:1: {print
awk: cmd. line:1:        ^ unexpected newline or end of string

As an aside, I understand using cat is not as efficient as awk by itself, but very often I build one liners like this using cat, trimming output up as I go. If I were to script this out, I would consider a more efficient method, but if I'm generating a one-time report, it makes no difference to me.

Upvotes: 0

Chris Seymour
Chris Seymour

Reputation: 85795

If you have GNU grep:

$ grep -Po "(?<=')[^',]+(?=')" file | xargs -n2
DB_USER abc_12345
DB_PASSWORD abc12345
DB_USER def_34589
DB_PASSWORD def34589

This greps for anything inside single quotes that isn't a ' or , then xargs just group the matches two per line.

P.S never store passwords in plaintext!

Upvotes: 1

Jean-Pierre
Jean-Pierre

Reputation: 171

awk -F\' '{print $2,$4}' file.txt

Upvotes: 9

Pankrates
Pankrates

Reputation: 3094

You can use the record separator flag -F of awk and separate at '

$> string="define('DB_USER', 'abc_12345');"
$> echo $string | awk -F"'" '{print $2}'
DB_USER
$> echo $string | awk -F"'" '{print $4}'
abc_12345

Or just simply pass the file

awk -F"'" '{print $2 $4}' file.txt

Upvotes: 3

Related Questions