Create aggregation function to report most popular in Google DataTable -

i have google datatable looks this:
var data = new google.visualization.datatable(); data.addcolumn('string', 'city'); data.addcolumn('string', 'model');
data.addcolumn('number', 'sold'); data.addrows([ ["melbourne","ford",10], ["perth","ford",2], ["melbourne","ford",7], ["melbourne","holden",25], ["perth","holden",25], ["melbourne","holden",12], ["melbourne","ford",11] ]);

what group city , report model highest cumulative sold value city.

the result returned :

city model melbourne holden perth ford

from reading of google visualisation api possible write custom aggregation function beyond limited experience.

any assistance appreciated.

this require double grouping: first time total sold per model per city, , second time popular model.

var g1 = google.visualization.data.group(data, [0, 1], [{     type: 'number',     label: 'total sold',     column: 2,     aggregation: google.visualization.data.sum }]); // create view merge columns 1 , 2 grouping below var view = new google.visualization.dataview(g1); view.setcolumns([0, {     type: 'string',     label: 'model qty sold',     calc: function (dt, row) {         var o = {model: dt.getvalue(row, 1), sold: dt.getvalue(row, 2)};         return json.stringify(o);     } }]); var g2 = google.visualization.data.group(data, [0], [{     type: 'string'     label: 'most popular'     aggregation: function (values) {         var max = null, ret = null;         (var = 0; < values; i++) {             var o = json.parse(values[i]);             if (max == null || o.sold > max) {                 max = o.sold;                 ret = o.model;             }         }         return ret;     } }]); 

g2 should contain model highest sales grouped city. if there tie between 2 models, return first model in values array (likely values ordered model name, not guaranteed).