Unicode - ANSI Interface question

[Former Member]

An R3 4.7 interface in a non-unicode system outputs a file in ANSI for a 3rd party system.

The R3 4.7 system is then upgraded to Unicode (Oracle UTF-8 database). The 3rd party system still expects the data in ANSI format.

Will this Interface still work? It can be assumed that the data will be LATIN1 characters only.

[sridhar_k1]

Is the interface program using OPEN DATASET to output the file?

If ENCODING NON-UNICODE option used in open dataset statement, it works

if ENCODEING DEFAULT used, it outputs unicode(UTF-8) file.

If GUI_DOENLOAD used to output file, the interface still works.

Regards

Sridhar

[Former Member]

You can used OPEN Dataset with Encoding option if you want the transfer the output file on the application server or else you can use GUI_DOWNLOAD function module for file on the presentation server.

I think ther should not be any issue,the interface would work.

[Former Member]

I have a number of abaps exporting data via open dataset encoding default (i.e. UTF-8). Also some using CL_FRONTEND_SERVICES GUI UPLOAD. And also IDOCS too.

I have done some searching and it appears as though as long as the characters are all LATIN1 (which all the existing interfaces are) then the UTF-8 output will be the same as ASCI or ANSI because code points 0 - 127 are single byte in UTF-8 and stored the same.

Has anyone else had any experience of this?

Thanks

[sridhar_k1]

That's right, UTF-8 uses one byte between 0 - 255. The files will be same as ANSI as long as sap don't add byte order mark EFBBBF (standard for UTF-8 files) at the begining of the file.

To verify, open a sap created file in a hex viewer, if first three bytes are EFBBBF, then they should be removed.

Regards

Sridhar

Message was edited by: Sridhar K

[Former Member]

Thanks Sridhar, I will open a selection of files and investigate further.

Do you know what format the IDOCS will be in from a Unicode system? Will there be any change?

Thanks

[sridhar_k1]

You can select the file encoding in file port definition(Tx: WE21).

If you check unicode, idocs written to file port are UTF-8 else non unicode.

I've opened SAP UTF-8 file created from non UC system in hex viewer, there's no byte order mark.

Regards

Sridhar