Stream API tutorial

This is the tutorial on Stream API and functional programming techniques

Pre-conditions ❗

You're supposed to be familiar with OOP, have basic knowledge of JDK, and be able to write Java code.

Related exercises 💪

• Account analytics
• File Reader
• File Stats

See also 👇

• Tutorial on Lambdas
• Tutorial on Optional

Stream API provide an ability to process sequences of data elements in a declarative way and simplify the task of performing operations in parallel

The simplest example is the task of filtering a collection. Using imperative old-style approach we specify HOW the task should be done. E.g. dealing with iteration, and storing each element in a new ArrayList. This way of processing is also called external iteration. Stream API and it's declarative approach allows us to specify WHAT should be done, without actually dealing with iteration and elements. This approach is also called internal iteration.

Imperative style:

        List<Account> gmailAccounts = new ArrayList<>();
        for (Account account : accounts) {
            if (account.getEmail().endsWith("@gmail.com")) {
                gmailAccounts.add(account);
            }
        }

Declarative style using Stream API:

        List<Account> gmailAccounts = accounts.stream()
                .filter(a -> a.getEmail().endsWith("@gmail"))
                .collect(toList());

Best practices

• use streams wheneve it can make the code clear and concise
• use clear names for lambda parameters in stream pipelines
• use helper methods for better readability
• be carefull using a stream of chars
• prefer pure functions for stream operations, avoid side-effect function
• always use static import for better readability
• always use joining() collector for CharSequence elements to avoid performance issues
• prefer collections and not stream as method return-type
• use parallel stream only when it's critical, when you know that it helps, and always measure it using real data
• ALWAYS USE SPECIAL CLASSES FOR PRIMITIVES. (E.g. IntStream instead of Stream<Integer>)