Hi,

we recently upgraded to the latest DOA version. In one of our applications we link to a view that reads data from an other database.

In the older verion this worked ok. In the new version we get the error: expecting: string actual: widestring on a varchar2 field.

Is there a way to force this dataset to stringfields in stead of widestring. (Like it used to work?)

kind regards

Ruud

Is this a Unicode database? If so, you can set Oracle.NoUnicodeSupport to True to force Unicode data conversion to the local character set. This way you omit the widestring fields.

I have the same problem. Where do I set Oracle.NoUnicodeSupport to True?

At run-time during initialization of your application.

Thanks. That worked it out

Hello.

I would like to use this topic to know more about the DOA conversion by setting NoUnicodeSupport. I made a short test with a english and portuguese version of Windows, both seems to work with our common characters with accent.

There is a single mapping for all western languages? I'm specifically interested in english, spanish and portuguese.

With the Oracle XE 11.2, there is a fixed charset of AL32UTF8, so this is the only reason at the moment to deal with Unicode. I could be using ForceWideStringFields and changing all the TStringField - but with many hundred of thousands lines of code, it doesn't worth right now