If we want to get two types of information from a Stream of objects we can consume the Stream twice and collect the results. But that is not very efficient, especially when the stream has a lot of objects. Since Java 12 we can use the teeing method of the java.util.stream.Collectors class to get multiple results while consuming the stream of objects only once. The teeing method takes two collectors as argument each returning a separate result for the stream items. As third argument we must pass a function that will merge the results of the two collectors into a new object.

In the following code we have two example use cases that use the teeing method to get multiple results while consuming a Stream of objects only one time:

In Java we can merge a key/value pair into a Map with the merge method. The first parameter is the key, the second the value and the third parameter of the merge method is a remapping function that is applied when the key is already present in the Map instance. The remapping function has the value of the key in the original Map and the new value. We can define in the function what the resulting value should be. If we return null the key is ignored.

If we want to merge multiple Map instances we can use the Stream API. We want to convert the Map instances to a stream of Map.Entry instances which we then turn into a new Map instance with the toMap method from the class Collectors. The toMap method also takes a remapping function when there is a duplicate key. The function defines what the new value is based on the two values of the duplicate key that was encountered. We can choose to simply ignore one of the values and return the other value. But we can also do some computations in this function, for example creating a new value using both values.

In Java we can use the iterate method of the Stream class to create an unbounded stream based on function invocations. We pass to the iterate method an initial value and a function that can be applied to the value. The first element in the unbounded stream is the initial value, the next element is the result of the function invocation with as argument the value from the previous element and this continues for each new element. Suppose we have a function expressed as lambda expression i → i + 2. When we use this lambda expression with the iterate method and a initial value of 1 we get a stream of 1, 1 → 1 + 2, 3 → 3 + 2, …​.

As we get an unbounded stream we must for example use limit to get the values we want from the stream. But we can also use an extra argument for the iterate method that is a Predicate definition. The iterate method will provide elements as long as the result of the Predicate is true. This way we the result of the iterate method is a bounded stream.

In Java we can use the generate method of the Stream class to create an infinite stream of values. The values are coming from a Supplier instance we pass as argument to the generate method. The Supplier instance usually will be a lambda expression. To give back a fixed value we simply implement a Supplier that returns the value. We can also have different values when we use a method that returns a different value on each invocation, for example the randomUUID method of the UUID class. When we use such a method we can create the Supplier as method reference: UUID::randomUUID.

The generate method returns an unbounded stream. We must use methods like limit and takeWhile to get a bounded stream again. We must use findFirst or findAny to terminate the unbounded stream and get a value.