If you sort an index alphabetically, strings are ordered by the lexicographical order of their first 50 Unicode characters with the exception of some normalized characters.
The alphabetical sort will reflect the normalized version of these special characters rather than their unicode value.
For example, ⑲ is considered as being 19. This allows end users to search for 19 and find ⑲. This may break unicode order in specific cases.
In a strict unicode sort, ①大 would come before ⑲愛 because ① has a lower unicode value than ⑲. Because Algolia normalizes ⑲愛 to 19愛, it ends up being alphabetically "lower" than 1大 (the normalized version of①大) since 9 is lower than 大.
Please note that Algolia strongly discourages implementing an alphabetical sort as it devalues and essentially disables Algolia’s tie-breaking algorithm.