How do ranges [a-z] and negation [^abc] work in a Java character class, and how do you include a literal hyphen or caret?
answer
- Hyphen between two chars = range by code point
- Leading caret = negate the whole set
- Hyphen first/last/escaped = literal dash
- Caret is literal unless it is first
- [A-z] trap includes [ \ ] ^ _ `; negated class matches newlines
basics
~20 s[a-z] matches any one character whose code falls between 'a' and 'z'. [^abc] matches any one character that is NOT a, b, or c. Put a hyphen first or last (or escape it) to mean a literal dash.
solid answer
~40 sA range like [a-z] matches one character whose code point lies between the two endpoints inclusively, ordered by Unicode value, so [0-9] is digits and [A-Za-z] is ASCII letters. A leading caret negates the whole class: [^abc] matches any single character except a, b, or c, including newlines unless you constrain it. The hyphen is only a range operator between two characters; placed first ([-abc]), last ([abc-]), or escaped (\-) it is a literal dash. The caret is only special as the very first character; elsewhere ([a^]) it is a literal caret. Be careful with ranges that span case or punctuation, e.g. [A-z] accidentally includes the six characters between 'Z' and 'a' ([, \, ], ^, _, `).
go deeper
Knows [a-z] is a range and [^...] negates; can match digits with [0-9].
Correctly places a literal hyphen/caret, knows ranges order by code point, and avoids the [A-z] trap.
Explains negation includes line terminators, validates endpoints, and chooses ranges vs predefined classes deliberately.
Considers Unicode-correctness of ASCII ranges, input-validation security implications of overly broad negated classes, and readability/maintenance trade-offs.
## Recap: a class matches one character A character class `[...]` matches exactly one input character that is a member of the described set. Ranges and negation are two ways to describe that set more compactly. ## Ranges with the hyphen Writing `[a-z]` means "any one character whose **code point** is between `a` and `z` inclusive". A **code point** is the numeric value Unicode assigns to a character; characters are ordered by that number. So: - `[0-9]` = the ten ASCII digits. - `[a-z]` = lowercase ASCII letters; `[A-Z]` = uppercase. - `[A-Za-z]` = combine two ranges in one class (you can list several ranges and loose characters together, e.g. `[A-Za-z0-9_]`). The hyphen `-` is the **range operator** only when it sits **between two characters**. A reversed range like `[z-a]` is illegal and throws `PatternSyntaxException`. ### Range pitfalls Because ordering is by code point, a range can include surprises. `[A-z]` spans from `A` (65) to `z` (122) and therefore also matches the six punctuation characters between `Z` (90) and `a` (97): `[ \ ] ^ _ \``. Always pick endpoints in the same contiguous block, e.g. `[A-Za-z]` not `[A-z]`. ## Negation with the caret A caret `^` as the **very first** character inside the brackets **negates** the class: `[^abc]` matches any one character that is **not** `a`, `b`, or `c`. Negation also applies to ranges: `[^0-9]` is any non-digit. Important subtlety: a negated class matches **any** other character, **including line terminators** like `\n`, unless you add flags or further restrict the set. (The dot `.`, by contrast, excludes line terminators by default — see the dot question.) ## Including literal hyphen and caret - **Literal hyphen**: place it where it cannot form a range — first `[-abc]`, last `[abc-]`, or escape it `[a\-c]`. Otherwise `[a-c]` is read as a range. - **Literal caret**: it is only special as the first character. Anywhere else it is literal: `[a^b]` matches `a`, `^`, or `b`. To put a caret first as a literal, escape it: `[\^a]`. - **Literal closing bracket**: escape it `[a\]b]` (or place it first in some flavors; in Java escape it). ## Java source escaping Remember regex backslashes double in Java string literals: the regex `[a\-c]` is the Java literal `"[a\\-c]"`. Compile with `Pattern.compile(...)`. ## Deriving the answer Know three rules — (1) hyphen between two chars = range by code point, (2) leading caret = negate, (3) a metachar placed where it can't act becomes literal — and you can explain every example, including the `[A-z]` trap and negation matching newlines.
- Where can you place a hyphen so it is treated as a literal dash?First inside the class ([-abc]), last ([abc-]), or escaped ([a\-c]). In any of those positions it cannot form a range, so it matches a literal hyphen.
- Does [^abc] match a newline character?Yes. A negated character class matches any character not listed, including line terminators like \n. This differs from the dot, which excludes line terminators by default.
- Using [A-z] expecting only letters (it spans punctuation)
- Thinking the caret negates anywhere in the class, not just at the start
- Assuming a negated class excludes newlines like the dot does
- Writing a reversed range like [z-a] (throws PatternSyntaxException)