filter duplications by property name with java stream filter

Question

How can I use java streams to filter an object by the same porperty?

The HashSet tasks contains:

task=1, name=hello
task=2, name=hello

(I dont want to use equals/hashcode on "name")

Now I want to filter the tasks producing a new HashSet with unique name-property.

All objects having the same name should be treated as equal:

tasks.stream.filter(???);

I can solve that by creating a new list and with 2 for-loops by indexing next element and looking for the name if it was already put into the list. However, maybe there is an one liner for such problems? Maybe with stream.filter()?

What would be your output type. Please go through this and update your question with minimum reproducible example: https://stackoverflow.com/help/minimal-reproducible-example — Ravindra Ranwala, Aug 22 '19 at 06:17
Please state the expected output type in your problem statement clearly, otherwise it is misleading and different people may interpret it in a different manner. — Ravindra Ranwala, Aug 22 '19 at 07:03

Ravindra Ranwala · Accepted Answer · 2019-08-22T06:35:31.123

2

You may first create a map using the task name as the key and the relevant Task as the value. Ignore the duplicate Tasks merely by using a mergeFunction. Then get all the distinct tasks as a Set view.

Set<Task> distinctTasks = tasks.stream().collect(Collectors.collectingAndThen(
    Collectors.toMap(Task::getName, Function.identity(), 
        (a, b) -> a), m -> new HashSet<>(m.values())));

edited Aug 22 '19 at 06:35

answered Aug 22 '19 at 06:07

Ravindra Ranwala

20,744
6
45
63

nimo23 · Answer 2 · 2019-08-22T06:12:23.097

0

I find one solutions:

List<Task> tasksDistinct = tasks.stream()
                    .collect(Collectors.groupingBy(Task::getName))
                    .values()
                    .stream()
                    .flatMap(m -> m.stream())
                    .collect(Collectors.toList());

Are there any better solutions?

edited Aug 22 '19 at 06:12

answered Aug 22 '19 at 06:08

nimo23

5,170
10
46
75

1

where and what are you `filter`ing? – Naman Aug 22 '19 at 06:09
It is a indirect filtering by "Task::getName" and knowing that map key is unique. Do you know a solution with explicit `stream.filter()`? – nimo23 Aug 22 '19 at 06:11
But if you don't use that `Map` anymore, what are you achieving by `groupingBy`? You eventually have a list of all `Task`s, ordered possibly by the keys(`name`) iterated for the map. – Naman Aug 22 '19 at 06:13
`groupingBy` constructs the map with unique key. The origin task is a HashSet with duplicate names.. – nimo23 Aug 22 '19 at 06:15
Yes, `groupingBy` constructs the map with unique keys and *lists of all elements* as values. Your subsequent operation joins all of them back to a list containing all elements. You could avoid this by replacing `.flatMap(m -> m.stream())` with `.map(m -> m.get(0))`, to only use the first occurrence for every group, instead of all of them. But that’s less efficient than [the accepted answer](https://stackoverflow.com/a/57602947/2711488) which doesn’t keep the other occurrences in the first place. – Holger Aug 22 '19 at 09:44

filter duplications by property name with java stream filter

2 Answers2