Reputation: 187
I have a problem which I've solved, but I'm looking for a better solution.
I have a big list of (let's assume 500 000) entities
class Entity {
String key1;
String key2;
String value1;
String value2;
}
which I want to map into a multimap Map<String, Map<String, List<Entity>>>
based on key1 and key2 the way that the top level map has entity.key1
as key, and the inner map - entity.key2
.
Therefore I've constructed a mapper class to do such thing
class EntityMapper {
Map<String, Map<String, List<Entity>>> mapByKey1AndKey2(List<Entity> entities) {
Map<String, Map<String, List<Entity>>> accountMap = new HashMap<>();
entities.forEach(
entity -> key1Map.merge(entity.key1, newKey2Map(entity), this::mergeKey2Maps)
);
return key1Map;
}
Map<String, List<Entity>> newKey2Map(Entity entity) {
Map<String, List<Entity>> key2Map = new HashMap<>();
key2Map.put(entity.key2, new ArrayList<>(singletonList(entity)));
return key2Map;
}
Map<String, List<Delta>> mergeKey2Maps(Map<String, List<Entity>> oldMap, Map<String, List<Entity>> newMap) {
for (String key2 : newMap.keySet()) {
oldMap.merge(key2, newMap.get(key2), this::mergeLists);
}
return oldMap;
}
List<Entity> mergeLists(List<Entity> oldList, List<Entity> newList) {
oldList.addAll(newList);
return oldList;
}
}
Are there any disadvantages of this approach? Or is there any way to optimize it?
Disclaimer: I'm aware of the existence of composite key mapping, but it won't serve my purpose.
Upvotes: 0
Views: 60
Reputation: 28046
I think this should work:
Map<String, Map<String, List<Entity>> groupedEntities =
entities.stream()
.collect(Collectors.groupingBy(e -> e.key1,
Collectors.groupingBy(e -> e.key2)));
I first group the list by key1
, then I group the values of the resulting Map
by key2
.
Upvotes: 1