Question

I am writing a piece of code for a SSIS 2005 script component to read data from an informix database (where the database strings are stored as UTF8).

The output of this string needs to be loaded to a text stream (DT_TEXT), encoded using code page 1252 (ANSI – Latin I).

Here’s a simple example of what I am trying to accomplish (AllColumnsBuffer is the script component output buffer, ColumnText is the name of the DT_TEXT field I am loading).

        Dim s As String = "Testing,1,2,3" & System.Text.RegularExpressions.Regex.Unescape("\u4EB5")
        AllColumnsBuffer.AddRow()
        AllColumnsBuffer.ColumnText.AddBlobData(encoding.GetEncoding(1252).GetBytes(s))

I need to throw an error if the encoding finds characters that cannot be converted to 1252. It seems that now it just puts in a ? if a character in the source doesn’t exist. Is there any way to validate that the character exists in the target code page?

Answer 1

You could create a new encoding and then set the EncoderFallback property – either to your own fallback, or if an exception is good enough for you then you can use EncoderExceptionFallback. Your own fallback might (say) fail gracefully without an exception, but set a flag to tell you afterwards that it failed.