Actually, it seems that the best solution is to convert both strings to US7ASCII, as this eliminates all the accents. Therefore, I can do:
upper (convert (col_name, 'US7ASCII')), e.g. upper (convert ('% okopla%', 'US7ASCII'))
The only wrinkle I discovered is that the German eszett is turning into ?. Therefore, I will just search in my search term, and if it contains funny B, then I am going to use Gary NLS_UPPER ('große', 'NLS_SORT = XGerman'), otherwise I will just convert to ASCII. These are a bit of shreds, but we only need to cover English, French and German, so I think everything will be fine.
thanks for the help
source
share