Sort a list A to Z or Z to A. Ordering is numeric-aware, so item 2 lands before item 10 instead of after it the way a naive sort would put it.
A naive alphabetical sort compares strings character by character, so it puts item10 before item2. The character 1 sorts before 2 and the comparison stops there. That is correct for strings and wrong for anything a person reads. This tool sorts with a numeric-aware collator, which recognises digit runs inside text and compares them as numbers, so item2 lands before item10 and chapter 9 before chapter 11. The same collator handles accented characters properly, sorting é next to e rather than dumping every accented word at the end of the list where a byte-order sort would put it.
The unit you sort matters as much as the direction. Lines is the default and suits a stacked list. Words breaks the whole input on whitespace and sorts every token, which is how you alphabetise a paragraph of tags typed as one run. Comma-separated splits on commas and rejoins with a comma and a space, so a CSV-style tag string comes back sorted and still valid. Removing duplicates as part of the sort is a single toggle, and it applies before the ordering, so the count you end up with is the count of distinct items.
Because this one uses a numeric-aware collator. Digit runs inside text are compared as numbers, so 2 comes before 10 the way a reader expects.
No, that is a separate toggle. Turn on "remove duplicates" to dedupe and sort in one pass.
Next to their base letter, so é sorts with e rather than after z. That follows locale rules rather than raw byte order.