Stream: filter

Intermediate operations are lazy transformations returning a new Stream: filter keeps, map transforms, flatMap flattens, sorted buffers all, distinct uses HashSet, limit takes n, skip drops n, peek debugs.

filter(Predicate) keeps elements where test() returns true, is a lazy stateless operation backed by the Spliterator engine, and supports Predicate composition with and/or/negate.

Lambdas capture a copy of local variables which must be effectively final; use AtomicInteger for thread-safe mutable counters or final int[] for sequential-only workarounds.

anyMatch returns true on the first matching element, allMatch returns false on the first non-matching element, noneMatch returns false on the first matching element, and all three short-circuit.

takeWhile takes the leading prefix of elements where the predicate is true and stops at the first false. dropWhile skips that prefix and includes everything after. filter tests every element.

`distinct()` removes duplicates by `equals`/`hashCode`; to dedupe by a key, track seen keys in a `Set` or collect into a `Map` keyed by that field and keep the first value.

`anyMatch` = at least one, `allMatch` = every one, `noneMatch` = not a single one; all take a `Predicate`, short-circuit, and return a boolean.

Chain `Optional.flatMap`/`map`/`filter` to walk nested optional data safely; use `flatMap` when the step returns an `Optional`, `map` when it returns a plain value.

`findFirst` is deterministic by encounter order; `findAny` returns any match and is cheaper in parallel.

Filter on endsWith("@gmail.com") including the @ sign to avoid false matches like notgmail.com, and guard against null with short-circuit evaluation.

filter(Predicate) keeps elements that pass a boolean test and discards the rest, and for salary comparisons you must use BigDecimal.compareTo to avoid floating-point errors.

Use logical AND inside one predicate (or chain two filters) so an element survives only when both the department and join-year conditions are true.

Chain three conditions with AND so an element passes only when gender is male, age is over 30, and department is Sales, all simultaneously true.

An inclusive range filter keeps values where salary is at least the lower bound and at most the upper bound, endpoints included, using BigDecimal.compareTo for precision.

Keep only names that are not null and not blank by checking non-null first (short-circuit), then non-empty, using isBlank() on Java 11+ or trim().isEmpty() on Java 8.

Stream the map's entrySet, filter by value, and collect back into a map with toMap to produce a filtered copy.

Count each character with groupingBy and counting, then filter entries whose count exceeds one to find the characters that appear more than once.

Stream the characters with chars(), filter out whitespace with !Character.isWhitespace, map each code point to a one-character string, and join with no separator.

filter(s -> s.matches("[aeiouAEIOU]+")) keeps only strings composed entirely of vowels.