A stream builds a pipeline of lazy intermediate operations that does nothing until a terminal operation runs, and then elements flow through in one pass, which is why filter before map matters. Parallel streams split the source across the shared common ForkJoinPool. Use them only for large, cheaply splittable, CPU bound work with no shared mutable state, and never for blocking I/O, since you would tie up a pool the whole application shares.
Why interviewers ask this
The interviewer wants to know if you understand laziness, statelessness and where parallel streams go wrong, because misuse is common and hurts unrelated code through the shared pool. They are also listening for side effect free lambdas and correct collector use. It signals whether you write streams because they are clearer or because they look modern.
How to structure your answer
- Explain laziness and the single pass at the terminal operation.
- State the requirements for correct parallel execution.
- Warn about the shared common pool and blocking calls.
- Say how you decide: measure, do not assume.
Example answer
Intermediate operations just build a pipeline; nothing happens until a terminal operation pulls elements through, and each element is pushed through the whole chain rather than the collection being traversed once per operation. That is why ordering matters, filtering early does less work, and why short circuiting operations like findFirst can stop the whole source early. For parallel streams I am cautious. They split the source and run on the common ForkJoinPool, which is shared by the entire JVM, so a slow parallel stream in one request can delay everything else using it. That rules out any blocking work: I have seen a parallel stream making HTTP calls starve every other parallel stream in the process. Where they genuinely help is a big in memory computation over an array or an ArrayList, which split evenly, with no shared mutable state and no ordering dependency. Even then I benchmark, because the split and merge overhead often eats the gain on collections of a few thousand elements. If I need a dedicated pool I run the stream inside my own ForkJoinPool rather than the common one.
Walking into this interview soon? GhostPilot listens to your live call, spots the question the moment it is asked, and puts a structured answer on your screen in real time. Try it on your next mock, or grab a $29 Session Pass, no subscription, for the real thing.
See how it worksFollow-up questions to expect
- Why are LinkedList and streams from an iterator poor candidates for parallelism?
- What is the difference between reduce and collect?
- How do stateful lambdas break a parallel stream?
Related java developer questions
Your interviewer will ask their own version of this. Paste your actual job description into the free Question Predictor and get the 20 questions that role is most likely to ask, with what each one is really probing.
Predict my questions