extract xml array value in mysql or sas

Question

Can anyone please tell me how to extract this kind of data :

[{"number":"8457215152","type":"Cell","state":"LA","country":"US","tz":"CT","zip":"70546","msa":"0"},{"number":"4363685555","type":"Cell","state":"LA","country":"US","tz":"CT","zip":"70546","msa":"0"}]

I would like to have result to have something like this for this id

id number type state country tx zip msa 1 845... 1 436...

My problem is some id have more than two numbers ( this id have only 2 number ) I usually am able to use extractvalue function in mysql , but in this case , i'm at the end of my rope.

Thanks

Answer 1

    data work.parsed;
    infile cards;
    input;

    length line_str $32000 rec_str $800 number type state country tx zip msa $100 elemname $32;

    line_str = compress(_infile_, '"'); /* remove quotes */
    line_str = translate(line_str, ':', ','); /* make : a key:value separator */

    keep id number type state country tx zip msa;
    id = _N_;
    rec_count=countc(line_str, '{');

    array  elem {*} $ number type state country tx zip msa;/* order is important */

    put rec_count=;
    do r=1 to rec_count;
        if r = 1 then rec_start=3;
            else rec_start = rec_end + 4;
        rec_end = findc(line_str, '}', rec_start) - 1;

        rec_str=substr(line_str, rec_start, rec_end - rec_start + 1);

        do i=1 to dim(elem);
            elemname = vname(elem(i));
            elem(i)= scan(rec_str, i * 2, ':');/* this way relying on all elements provided in record in expected order */
            if findc(elem(i), '}') > 0 then elem(i) = substr(elem(i), 1, findc(elem(i), '}') - 1);
        end;
        output;
    end;
    cards;
    [{"number":"8457215152","type":"Cell","state":"LA","country":"US","tz":"CT","zip":"70546","msa":"0"},{"number":"4363685555","type":"Cell","state":"LA","country":"US","tz":"CT","zip":"70546","msa":"2"},{"number":"33333","type":"Cell","state":"CA","country":"US","tz":"CT","zip":"33333","msa":"3"}]
    ;
    run;

Surely this has some assumptions on what data looks like. HTH Vasja