Another dumb question

[jbv at souslelogo.com]

Hi again list,
I am processing large xml files (from 190 Mb to 1.6 Gb)
in order to extract nodes contents to be inserted in a DB,
and then displayed in various types of web pages or svg
graphs.
Because of the files size, I proceed by successive chunks
of 1000 lines.
These data include a small portion of japanese, russian
and arabic names & addresses, and the purpose is to keep
everything identical all the way from xml extraction to DB
to html & svg display.
So I use "set the useUnicode to true" at the beginning of
my main LC script, but how can I be sure that everything
will work as expected ? When I save temporary files of
the content of each successive portion of xml data, even
in utf8, japanese chars are replaced by strings of garbage...

Any help & advice will be much apreciated.
Thanks in advance.
jbv