Problem with get data using sed command

Hi,

I am using the following command(sed) to get the key/value pair from the string

String="{ "test":"test message", "testmessage":"subscription is active, charge successfully} " }"
status=$( echo $String | sed -e 's/^.*\("testmessage":[^,]*\).*$/\1/')
echo $status

i am getting this output :

"testmessage":"subscription is active

Expected output:

"testmessage":"subscription is active, charge successfully"

Please Suggest me,
Regards,
Nanthagopal A

Have ...

"test":"test message", "testmessage":"subscription is active, charge successfully"

Want ...

"testmessage":"subscription is active, charge successfully"

Try this modification of your sed ...

sed -e 's/^.*\("testmessage":[^,]*\)\(.*\)/\1\2/'

If you always want all characters following the first comma, try cut instead of sed ...

cut -d"," -f2-

This is less complicated but has the minor disadvantage of delivering the desired string with a leading blank.

Daniel B. Martin

Shouldn't the assignment statement use single quotes:

String='{ "test":"test message", "testmessage":"subscription is active, charge successfully" }'

Without that change I get an error (line 2 is the assignment, line 1 being the #!):

./keyval.sh: line 2: message, testmessage:subscription: command not found

Originally Posted by danielbmartin Have ... "test":"test message", "testmessage":"subscription is active, charge successfully"

The initial post is a bit mangled, but it seems to me that the curly braces were intended to be part of the string.

I would try looking for the closing quote:

sed -e 's/^.*\("testmessage":"[^"]*"\).*$/\1/'
## or
grep -o '"testmessage":"[^"]*"'

I agree with ntubski ... please carefully re-write the string as its current incarnation cannot be solved using the suggestions available due to extra curly brace.