Question

I have some xml encoded as UTF-8 and I want to write this to a Text field in SQL Server. UTF-8 is byte compatible with Text so it should be able to do this and then read out the xml later still encoded as utf-8.

However special characters such as ÄÅÖ, which are multi-byte in UTF-8 get changed on the way.

I have code like this:

byte[] myXML = ...

SqlCommand _MyCommand = new SqlCommand(storeProcedureName, pmiDB.GetADOConnection());
_MyCommand.CommandType = CommandType.StoredProcedure;
_MyCommand.Parameters.Add("xmlText", SqlDbType.Text);
_MyCommand.Parameters["xmlText"].Value = Encoding.UTF8.GetString(myXML);
_MyCommand.ExecuteNonQuery();

My guess is that changing the xml byte array to string changes the special characters to UTF-16 characters which are then changed again to the Latin1. And Latin1 ÖÄÅ are not the same as UTF-8 ÖÄÅ.

How can I write the UTF-8 xml bytes to the Text field without them getting changed?

Answer 1

The solution that I got to work was to change the Stored Procedure so that the myXml parameter was Varbinary(Max), which allowed me to pass in the byte array. Then in the SP I Cast the Varbinary(max) to Varchar(max). This preserves the bytes as required for UTF-8

SET myXMLText = CAST(myXMLBinary as VARCHAR(MAX))