Good morning, We are planning a migration to unicode. Actually we are using Oraclce9.2.0.6 and we also want to move to 10.2 First of all, regarding the Unicode conversion, here are my questions: Based on the report generated by csscan, we will have some truncated data. To resolve that, we will increase the size of the specified varchar2. But what about the other varchar2 fields ? Question 1: Should I increase every varchar2. Example: in a varchar2(5), before the migration, I can insert 5 characters. After the migration, I might insert only 3. But I still want to insert 5, so should I increase the size to vachar2(10) ? Even if it was not specified in the report from csscan ? Question 2: After the migration to AL32UTF8, when we have a field with varchar2(4000 bytes), only 1000 characters can be inserted. What should we do with our actual fields with more then 1000 characters in it ? Should we convert them in CLOB ? Question 3: Why the use of NLS_LENGTH_SEMANTICS=char will help us ? We need some clarifications ? There is not a lot of documentation. Can someone can give me some pros/cons regarding the use of this parameter ? Thank you very much Luc -- Luc Demanche Oracle DBA (514) 867-9977