Constantine Yannakopoulos wrote > On Mon, May 12, 2014 at 4:46 PM, Jan Slodicka <
> jano@ > > wrote: > I understand that it is difficult to find the least greater character of a > given character if you are unaware of the inner workings of a collation, > but maybe finding a consistent upper limit for all characters in all > possible collations is not impossible? > > -- Constantine > _______________________________________________ > sqlite-users mailing list > sqlite-users@ > http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users I could be wrong when I refused this idea. U+10FFFD (represented as surrogate pair "\uDBFF\uDFFD") may be that character. So one could replace "LIKE 'xxx%'" by "BETWEEN('xxx', 'xxx' + '\uDBFF\uDFFD'). U+10FFFD is the highest codepoint from the Unicode Private Use Area (PUA). Unicode standards do not specify anything for this area. So vendors and/or apps (MS Word for example) can define here their own characters incl. their properties (font, sorting etc.). I saw somewhere that somebody placed here Japanese Kanji characters, for example. msdn statement for CompareString: "All characters in the PUA are sorted after all other Unicode characters. Within the area, characters are sorted in numerical order." Hence Windows is OK. I could not find any such notion for UCI library, but when I tested their interactive demo, character U+10FFFD was really sorted at the end. (Even with Japanese, Hindi etc.) That might mean that at least the standard setup of the UCI library sorts U+10FFFD in the expected way. Of course, this is no proof - UCI library can be customized. After all, it seems that U+10FFFD is a good choice, at least for a vast majority of cases. BTW, my tests showed that the speed improvement of the LIKE optimization is in the range 10x-100x. Any comments are welcome. -- View this message in context: http://sqlite.1065341.n5.nabble.com/LIKE-operator-and-collations-tp73789p75643.html Sent from the SQLite mailing list archive at Nabble.com. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
