Sort Lines

What this tool does

Sorts a list of lines five different ways: alphabetically, in natural order, by the number each line starts with, by line length, or shuffled into a random order. Each mode can be reversed, made case-sensitive, told to ignore indentation, and told to throw away blank lines first.

Natural order is the one worth the visit. Plain alphabetical sorting compares text character by character, so file10 lands before file2 — because 1 precedes 2. Natural order reads runs of digits as numbers and puts them where a human expects.

Common uses

• Alphabetising a list of imports, environment variables, dependencies or CSS custom properties so the next diff is readable.
• Ordering filenames or version strings correctly — v1.9.0 before v1.10.0, which alphabetical sorting gets backwards.
• Ranking log lines by a leading duration or count with numeric mode.
• Sorting by length to find the outlier — the one 300-character row in a file of 40-character rows is almost always the malformed one.
• Shuffling a list to pick a random order for a rota, a test fixture or a draw.

A short example

Alphabetical order gives:

file1.txt
file10.txt
file2.txt
file20.txt
file3.txt

Natural order gives what you meant:

file1.txt
file2.txt
file3.txt
file10.txt
file20.txt

How each mode compares

Alphabetical compares by Unicode code unit, lowercased unless "case sensitive" is on. With case sensitivity enabled, every uppercase ASCII letter sorts before every lowercase one, so Zebra precedes apple — that is how sort behaves in the C locale, and it surprises people every time.

Natural walks both strings as alternating runs of digits and non-digits. Digit runs compare by value, and because they are compared by length-then-text rather than by parsing to a float, IDs longer than 15 digits still order correctly — a real problem for any implementation that reaches for parseInt. Zero-padded forms are kept deterministic: img001 groups just ahead of img1.

Numeric reads a number from the start of each line, understanding a leading sign, a decimal point and thousands separators, and sends lines with no number to the bottom in their original order. Length counts UTF-16 units. Shuffle uses a Fisher-Yates pass, which is uniform — unlike the sort(() => Math.random() - 0.5) trick, which is measurably biased.

Worth knowing

Every sort here is stable, so lines that compare equal keep their original relative order. That lets you chain passes: sort by one key, then by another, and the first ordering survives inside each group of the second. "Use result as input" makes that a single click.

Sorting is done on the line's text, not on a locale collation. ä therefore sorts after z rather than beside a, which is right for identifiers and wrong for a German phone book. For human-language lists in a non-English locale, your spreadsheet's locale-aware sort is the better tool.