Problem with saving chinese character into oracle using NHibernate

Question

I am trying to save chinese character into oracle DB using Nhibernate thru C# and the character saved always end up with some rubbish character.

I have done the following:

1. Hbm mapping for the field is declared as type AnsiString and column type as sql-type="nvarchar". Have also tried String and nvarchar2 too.
2. Database column data type is NVARCHAR2.
3. Registry, environment variable and database charset NLS_LANG is set to UTF8.

I am able to read the chinese character input thru sql developer. But when I try to retrieve them (the character display correctly on screen) and attempt to save them back into the database without any changes, the data saved became "�����".

Anyone have any clue what have I miss out here?

Answer 1

I would check what the data really is on the .NET side. Don't trust what's displayed on a console - look at the UTF-16 code points.

See my debugging unicode article for the general approach.

I seem to remember having some problems in Oracle before which were only fixed by setting the database encoding and then rebuilding the database - I couldn't apply it to an existing one. However, the details are in the dim and distant past...