Line Sorter

Sorting a list of lines sounds trivial — alphabetical means A before B, right? It does, except when your list contains numbers, mixed cases, or human-meaningful patterns. file1.txt, file2.txt, file10.txt alphabetically becomes file1.txt, file10.txt, file2.txt because "1" < "2" character-wise. That's correct by the rules of lexicographic comparison and almost never what you want. Natural sort — treating runs of digits as numbers — produces the intuitive ordering and is the right default for filenames, version numbers, and most human-curated lists.

When to use each mode

Alphabetical (strict lexicographic) is right when you have no embedded numbers and care about exact character order — e.g. a list of names, words, slugs, or labels. Natural is right when your lines contain numbers that should sort numerically — file paths, version tags, sortable identifiers, episode titles. Length is right when you're spotting outliers — the longest URL in a log, the shortest acceptable password from a generated list, the most concise variant of a wording. Shuffle is right when you want randomised order — e.g. randomising a list of names for a draw, picking a non-obvious anonymisation order, generating a randomised test order.

The dedupe option is more useful than it looks

A surprising amount of real-world list work is "deduplicate this thing" — extracted email addresses, log paths, queried identifiers, URLs from a crawl. The tool's dedupe pass is exact-match (line-by-line) and respects the case-insensitive toggle, so [email protected] and [email protected] collapse to one entry when CI is on. For more sophisticated email-deduplication (Gmail dot variants, plus-tag aliases), use the dedicated Email Normalizer tool.

ASEAN-specific quirks

Localised sort orders matter when handling Asian-language content. Singaporean and Malaysian English content sorts the same as global English — straightforward. Indonesian and Filipino content is similar but watch for embedded diacritics (café should sort with cafe, which JavaScript's localeCompare() handles by default in this tool). Chinese characters sort by Unicode code-point order, which roughly tracks frequency of use — not pinyin-alphabetical, not stroke-count. If you need pinyin sort, pre-convert via the Chinese Converter first.

Why shuffle uses crypto.getRandomValues

The shuffle mode runs Fisher-Yates with window.crypto.getRandomValues() as the entropy source, so the resulting order is statistically uniform across all possible permutations — no clock-seeded predictability. For a list of 100 items, that's 100! ≈ 10^158 possible orderings. Math.random()-based shuffles can only reach a few billion of those (limited by the PRNG's internal state), which sometimes matters for cryptographic-grade randomness. For everyday list shuffling, the difference is invisible; we use the secure source anyway because it's the right default.

Composing operations with Output → Input

The Output → Input button enables piping the result back through another transform. Common chains: sort first, then dedupe to eliminate adjacent duplicates and rerun stats. Or: sort length-first, output the longest 10 manually, paste back, then alphabetise. The chaining cost is one button click and no re-paste — useful when you're building a list iteratively. Power users sometimes treat the tool as a personal data-cleaning pipeline: trim, then dedupe, then natural-sort, then output to a final destination. Each step is reversible (the input textarea keeps its previous value until you click Output → Input).