Quick sed question
 

I have a series of files in the format:

{"field(1)":"value(1)","field(2)":"value(2), ... ,"field(n)","value(n)"}

Which I'm trying to split using sed to have each field/value pair on on line:

sed -e '1s/^.//' -e '$s/.$//' -e 's/,/\n/g'

So far so simple, but some of the values might have commas in them, which means they'll be split up. Is there a way to change this so that only commas not within quotes will be replaced with \n? (e.g. make sed count number of " and if it finds a , after counting an odd number of " then ignore it?)

Thanks
Col

If it's exactly that, then you can replace runs of "," with "\n" and that will only match those commas between quotes rather than inside them.

Of course, if the data isn't that exact, then that may not work.

I never thought of that... Thanks!