In german wiktionary sortkeys of categories are stored in the database table categorylinks in field cl_sortkey in uppercase. Some time ago (i dont know when this happend) mediawiki software changed the behavior the character 'ɪ' Unicode Character 'LATIN LETTER SMALL CAPITAL I' (U+026A) is handled.
Before this change the character 'ɪ' was not converted to uppercase when stored into field cl_sortkey. After that software change the character 'ɪ' is converted to uppercase Unicode Character 'LATIN CAPITAL LETTER SMALL CAPITAL I' (U+A7AE).
It seems that it has been forgotten to patch the old values, which where entered before this software change occured. So now we have have a mixture of sortkey values in the database some of them contain U+026A and some U+A7AE.
This has confusing effects in viewing the category pages, because some entries dont appear in the list.
https://de.wiktionary.org/w/index.php?title=Kategorie:Reim_(Deutsch)&pagefrom=%C6%86%C9%AA%CC%AFt
The page https://de.wiktionary.org/wiki/Reim:Deutsch:-%C9%94%C9%AA%CC%AFt%C9%99 is missing in the list.
To verify the problem in the database make a select on database table categorylinks and compare the values generated for the field cl_sortkey. Some random examples not related to the links posted above:
cl_from cl_to cl_sortkey cl_timestamp cl_sortkey_prefix cl_collation cl_type 301735 Reim_(Deutsch) ɪNDƏT DEUTSCH:-ɪNDƏT 2013-01-21T12:21:58 ɪndət uppercase page 925336 Reim_(Deutsch) ꞮTVⱯ DEUTSCH:-ꞮTVⱯ 2019-04-21T12:56:30 ɪtvɐ uppercase page
The cl_timestamp of the last cl_sortkey value which contains an 'ɪ' might help to find the time when this behavior changed.
Unfortunately i can not say which categories are effected, it might be all. So to correct this i think it is the best to regenerate all sortkeys of all categories. Or just those which contain an 'ɪ', but i dont know if this is the only character which is effected by the software change.
And even more unfortunately this would have to be done on all other wikis, if not done yet.