Sort lines alphabetically or in natural order

Paste the lines into the box and pick A to Z — the sorted list appears underneath straight away, with no button to press and no signup. Everything is optional after that: ignoring capitals is on by default, so a mixed-case list comes out the way a person would file it rather than with all the capitals bunched at the top.

Because it sorted by character number instead of by alphabet. In the underlying character table every capital letter comes before every lowercase one — A is 65 and a is 97 — so a raw comparison files Zebra above apple and puts Smith between SMITH and smith. It is fast, it is what a programming language does by default, and it is almost never what a reader expects. This page uses proper collation instead, where a and A are the same letter at different strengths. The code point order setting is still available in the menu when you specifically want the raw version, and the strip under the box says which one you are looking at.

Because alphabetically it does: the comparison reaches the 1 and the 2 in the second position and 1 comes first, so nothing after that matters. This is correct alphabetical behaviour and almost never the desired one, which is why natural order exists — it spots runs of digits and compares them as numbers, so file2, file10 and file100 come out in the order a person would count them. Natural order is also what a file browser has been quietly doing since the mid-2000s, and it is the reason your downloads folder does not read 1, 10, 11, 2.

Not from a list written first name first — no sorter can, because nothing in the text says which word is the surname. Van der Berg, Maria Teresa and Kim Min-jun each have a different answer, and a guess would be wrong often enough to be useless. Rewrite the lines as Surname, First before sorting, which is what a library catalogue does and what makes the result checkable; a spreadsheet or a find-and-replace pass will do the rewriting in bulk. Once the lines start with the surname the sort is ordinary alphabetical.

Yes. When two lines compare as equal — which happens whenever capitals are being ignored, or two entries are genuinely identical — they come out in the order they went in, because the sort is stable and equal lines are settled by their original position. That guarantee has been part of the language standard since 2019 and it is what lets you sort twice to get a two-level order: sort by the secondary field first, then by the primary one, and the first sort survives inside each group of the second.

Only if you tick the box, and then it keeps the first copy of each line and drops the rest, reporting the number removed in the strip under the counter. Whether Apple and apple count as duplicates depends on the capitals setting, so the two switches work together. Sorting is a good moment to deduplicate because repeats end up adjacent and become obvious, but the two operations are separate on purpose — plenty of lists need the repeats kept.

Because å, ä and ö are letters in their own right at the end of the Swedish alphabet, not decorated versions of a and o. Sort Ängel in English and it files under A; sort it in Swedish and it comes after Zebra, which is what a Swedish reader will look for. The same divergence turns up all over: German telephone directories read ä as ae while ordinary German dictionaries file it under a, Czech treats ch as a single letter that follows h, Turkish keeps dotted and dotless i apart, and French breaks ties on accents by reading the word from the end. Ten alphabets are offered above, all supplied by the browser's own collation data.