Items i dont think you're gonna find via this post on PerlMonks.

• someone who is conversant with the Japanese script who will evaluate your data file. I would suggest you try with a local Perl Mongers group like Tokyo.pm
• a decent perl developer to help you write the code, if you say there are 30 files, with millions of lines in each.

Items for your reference..

• You have to take the support of CPAN modules like Unicode::Japanese.
• A related query posted earlier. To convert the unicode entity (症) to japanese character 在
• I'm sure monks can point you in the right direction, but for that can you please provide a more accurate dataset in your Original Post, along with some code (or pseudo code/design if you are new)