After testing, here is my best intersection approach.
Faster speed compared to pure HashSet Approach. HashSet and HashMap below has similar performance for arrays with more than 1 million records.
As for Java 8 Stream approach, speed is quite slow for array size larger then 10k.
Hope this can help.
public static List<String> hashMapIntersection(List<String> target, List<String> support) {
List<String> r = new ArrayList<String>();
Map<String, Integer> map = new HashMap<String, Integer>();
for (String s : support) {
map.put(s, 0);
}
for (String s : target) {
if (map.containsKey(s)) {
r.add(s);
}
}
return r;
}
public static List<String> hashSetIntersection(List<String> a, List<String> b) {
Long start = System.currentTimeMillis();
List<String> r = new ArrayList<String>();
Set<String> set = new HashSet<String>(b);
for (String s : a) {
if (set.contains(s)) {
r.add(s);
}
}
print("intersection:" + r.size() + "-" + String.valueOf(System.currentTimeMillis() - start));
return r;
}
public static void union(List<String> a, List<String> b) {
Long start = System.currentTimeMillis();
Set<String> r= new HashSet<String>(a);
r.addAll(b);
print("union:" + r.size() + "-" + String.valueOf(System.currentTimeMillis() - start));
}