Assuming I have the following:
var array =
[
{\"name\":\"Joe\", \"age\":17},
{\"name\":\"Bob\", \"age\":17},
{\"name\":\"Carl\
You could use a dictionary approach like this one. Basically you assign the value you want to be distinct as a key in the "dictionary" (here we use an array as an object to avoid dictionary-mode). If the key did not exist then you add that value as distinct.
Here is a working demo:
var array = [{"name":"Joe", "age":17}, {"name":"Bob", "age":17}, {"name":"Carl", "age": 35}];
var unique = [];
var distinct = [];
for( let i = 0; i < array.length; i++ ){
if( !unique[array[i].age]){
distinct.push(array[i].age);
unique[array[i].age] = 1;
}
}
var d = document.getElementById("d");
d.innerHTML = "" + distinct;
This will be O(n) where n is the number of objects in array and m is the number of unique values. There is no faster way than O(n) because you must inspect each value at least once.
The previous version of this used an object, and for in. These were minor in nature, and have since been minorly updated above. However, the reason for a seeming advance in performance between the two versions in the original jsperf was due to the data sample size being so small. Thus, the main comparison in the previous version was looking at the difference between the internal map and filter use versus the dictionary mode lookups.
I have updated the code above, as noted, however, I have also updated the jsperf to look through 1000 objects instead of 3. 3 overlooked many of the performance pitfalls involved (obsolete jsperf).
Performance
https://jsperf.com/filter-vs-dictionary-more-data When I ran this dictionary was 96% faster.