Am 07.06.2013 17:55, schrieb Dimitry Sibiryakov: > 07.06.2013 17:31, Leyne, Sean wrote: >> If an operator is comparing OCTETS against a string value, OCTETS >> value should be coerced into the charset of the other value. > > In which way? How code x'a0' should be "coerced" into UTF-8? As russian > "a" or western "รก"? > A long time ago I suggested to convert OCTETS into hexadecimal values, but > it was rejected.
I have no idea, how the x'##' should be treated, but it sure is quite irritating, reading in the docs that '[[:WHITESPACE:]]' would match some characters named by their ASCII-numbers - and it does using "ascii_char(##)" but doesn't using "x'##'" - we should have a hint there, at least... Moreover: The problem did arise in trying to find some text behind empty lines in a blob - that is charset ISO8859_1 and the connection settings are ISO8859_1 too. select iif(BlobField similar to '[[:WHITESPACE:]]*SomeTextToFind', 'T', 'F') from SomeTable would not return 'T', even though BlobField contains only CRLF in front of SomeTextToFind. If you need a more detailed example, I'll set one up. Thanks for the discussion, Thomas -- thomas.beckm...@thbeckmann.de ------------------------------------------------------------------------------ This SF.net email is sponsored by Windows: Build for Windows Store. http://p.sf.net/sfu/windows-dev2dev Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel
