>I think that my problem is in using LIKE expression for non-ascii strings. >Database encode is UTF-8. When table data in the "base" column (see my >first >message for structure) consists of english symbols (ascii) LIKE works >correct, but when I'm trying to execute it on strings consists of UTF8 >symbols non compatible with ascii it doesn't work.
As far as I understand it (please correct if this is wrong), SQLite will store/retrieve data "as supplied". If you hand it ANSI strings, you get them stored verbatim and can retrieve them verbatim. But and that's a big BUT without pun, if you use any function acting on data, just as LIKE, it reads UTF-* data (replace * by the encoding you used for creating the database) and _replace badly formed UTF-* characters by the Unicode u+FFFD (invalid codepoint) marker. From then on, the changed data no longer compares with the original string since some characters are destroyed in the process. I suppose indexing with anything else than the default BINARY collation is likely to produce erroneous results. The solution is to build a true UTF-8 (or UTF-16) database. The simplest way could be (again correction is welcome) to rebuild the base from the original data, once it is converted to valid UTF-* suitable for input. If your only data source now is the current base itself (input data no more available) you could try to dump the base with the command line utility and work from there. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
