dividing and aggregating json data

Question

I have some JSON data I need to divide then aggregate. Here is a sample of its format.

data2 = [{"startDate": 1396263600.0, "adId": 2483231759355, "endDate": 1401101940.0,        "impressions": 754831},
{"startDate": 1393851600.0, "adId": 2750329551133, "endDate": 1404212340.0, "impressions":    3947368},
{"startDate": 1401620400.0, "adId": 1311595275159, "endDate": 1404212340.0, "impressions": 630000}];

I want to uniformly divide up the impressions for each day in the date range. (ie if there are 100 impressions in a range of 10 days then each day will get 10 impressions). Then I want to aggregate the number of impressions on all the adIds for each day (ie if there are 2 ads on one day, then they get combined). So the final result will be a array of days and impressions. Not too familiar with javascript so not sure if there are any functions to make my life easier here. Any suggestions?

I plan on using this function to count number of days between dates:

function days_between(date1, date2) {

    // The number of milliseconds in one day
    var ONE_DAY = 1000 * 60 * 60 * 24;

    // Convert both dates to milliseconds
    var date1_ms = date1 * 1000;
    var date2_ms = date2 * 1000;

    // Calculate the difference in milliseconds
    var difference_ms = Math.abs(date1_ms - date2_ms);

    // Convert back to days and return
    return Math.round(difference_ms/ONE_DAY);

}

Answer 1

Are you thinking about something like this?

function days_between(date1, date2) {
    var date1_ms = new Date(date1*1000).getTime();
    var date2_ms = new Date(date2*1000).getTime();

    // Calculate the difference in milliseconds
    var difference_ms = Math.abs(date1_ms - date2_ms);

    // Convert back to days and return
    return Math.ceil(difference_ms / (1000 * 3600 * 24));
}


function aggregate(data) {
    result = [];
    for (var i = 0; i < data.length; i++) {
        var numOfDays = days_between( data[i].startDate, data[i].endDate );
        var adId = data[i].adId;
        var impressionPerDay = (data[i].impressions / numOfDays);
        for (var j = 0; j < numOfDays; j++) {
            var obj = {};
            obj.date = new Date(new Date(data[i].startDate*1000).getTime() + (j * (1000 * 3600 * 24))).getTime();
            obj.impressions = impressionPerDay;
            obj.adId = adId;
            result.push(obj);
        }
    }
    return result;
}

function aggregateAdIds(data) {
    var result = {};
    for (var i = 0; i < data.length; i++) {
        if (data[i].adId in result) result[data[i].adId].push({"date": data[i].date, "impressions": data[i].impressions});
        else result[data[i].adId] = [{"date": data[i].date, "impressions": data[i].impressions}];
    }
    return result;
}

var data = [{"startDate": 1396263600.0, "adId": 2483231759355, "endDate": 1401101940.0,        "impressions": 754831},{"startDate": 1393851600.0, "adId": 2750329551133, "endDate": 1404212340.0, "impressions":    3947368},{"startDate": 1401620400.0, "adId": 1311595275159, "endDate": 1404212340.0, "impressions": 630000}];


var final = aggregateAdIds(aggregate(data));

UPDATED

function days_between(date1, date2) {
    var date1_ms = new Date(date1*1000).getTime();
    var date2_ms = new Date(date2*1000).getTime();

    // Calculate the difference in milliseconds
    var difference_ms = Math.abs(date1_ms - date2_ms);

    // Convert back to days and return
    return Math.ceil(difference_ms / (1000 * 3600 * 24));
}


function aggregate(data) {
    var map = {};
    var result = [];
    for (var i = 0; i < data.length; i++) {
        var numOfDays = days_between( data[i].startDate, data[i].endDate );
        console.log('days : '+ numOfDays);
        var impressionPerDay = (data[i].impressions / numOfDays);
        for (var j = 0; j < numOfDays; j++) {
            var tempDate = new Date(data[i].startDate*1000);
            tempDate.setDate(tempDate.getDate() + j);
            var date = new Date(tempDate.getFullYear(), tempDate.getMonth(), tempDate.getDate(), 0, 0 ,0, 0);
            if (date.getTime() in map) map[date.getTime()] += impressionPerDay;
            else  map[date.getTime()] = impressionPerDay;
        }
    }
    for (var key in map) {
        result.push({'date' : Number(key), 'impressions' : map[key]});
    }
    result.sort(function(a,b){return a.date > b.date});
    return result;
}

var data = [{"startDate": 1396263600.0, "adId": 2483231759355, "endDate": 1401101940.0,        "impressions": 754831},{"startDate": 1393851600.0, "adId": 2750329551133, "endDate": 1404212340.0, "impressions":    3947368},{"startDate": 1401620400.0, "adId": 1311595275159, "endDate": 1404212340.0, "impressions": 630000}];


console.log(aggregate(data));

updated fiddle : link

Answer 2

So something like this? This gets the amount of impressions in each day into a javascript object, with each date in miliseconds as key (for example 1403953200000:21000 ). You can convert the miliseconds into a date if you want, and 21000 is the impressions on that day

var result={};

for(var i=0;i<data2.length;i++){
    var end = new Date(data2[i].endDate*1000);
    var start = new Date(data2[i].startDate*1000);
    var diff= Math.abs(end.getTime() - start.getTime());
    var diffDays = Math.ceil(diff / (1000 * 3600 * 24)); 

    for (; start <= end; start.setDate(start.getDate() + 1)) {
        var miliseconds=start.getTime();
        if(!result.hasOwnProperty(miliseconds)) result[miliseconds]=0;
        result[miliseconds]+=data2[i].impressions/diffDays;
    }

}

console.log(result)

Update:

function aggregate(data2) {
    var result={};
    var result_arr=[];    
    for(var i=0;i<data2.length;i++){
        var end = new Date(data2[i].endDate*1000);
        end.setHours(0);
        end.setMinutes(0);
        end.setSeconds(0);
        var start = new Date(data2[i].startDate*1000);
        start.setHours(0);
        start.setMinutes(0);
        start.setSeconds(0);        
        var diff= Math.abs(end.getTime() - start.getTime());
        var diffDays = Math.ceil(diff / (1000 * 3600 * 24)); 
        console.log(start.toString(),end.toString(), diffDays)
        for (; start <= end; start.setDate(start.getDate() + 1)) {
            var miliseconds=start.getTime();
            if(!result.hasOwnProperty(miliseconds)) result[miliseconds]=0;
            result[miliseconds]+=data2[i].impressions/diffDays;
        }
    }

    for(var key in result){
        result_arr.push({date:Number(key), impressions: result[key]});
    }
    result_arr.sort(function(a,b){return a.date > b.date})
    return result_arr;
}

Fiddle with my working solution: http://jsfiddle.net/juvian/6R7mu/14/