topic Re: Field Size differences in tables when upgrading to Unicode in Application Development and Automation Discussions

Field Size differences in tables when upgrading to Unicode

Former Member — Mon, 07 Aug 2006 15:36:55 GMT

WE have some programs that extract data from Tables in SAP such as USR06 and USR02. Since we upgraded soem of our systems to Unicode the programs are either not returnign data or appear to be misreading the fields. Any one have any idea on what changed with Unicode? I need to be able to help the developers understand what is going on. We have a 3rd party tool that is changing passwords on all of our systems but it is also used to retrieve this data. There developers can not figure this out.

Re: Field Size differences in tables when upgrading to Unicode

Former Member — Mon, 07 Aug 2006 18:37:19 GMT

It may be that your new unicode systems are using code version E for the password, but your 3rd party is tying to send a code version B hash to USR02.

Re: Field Size differences in tables when upgrading to Unicode

Wolfgang_Janzen — Mon, 21 Aug 2006 10:52:25 GMT

Most likely this problem is not related to Unicode but to the fact that the structure of the table (at least for USR02 I'm sure) has changed during the upgrade.

Please notice that the <b>internal data structures are subject of change without prior notification</b>. You should only use API functions to access data (read/write).

Regarding passwords:

please resist from implementing any direct table modifications - for audit and stability reasons. Again, there have been changes in the past (last recently: a major change in 7.0) and there might be more in the (near) future ... (I'm just working on one).

Regarding password synchronization:

I still don't understand why so many people believe in that approach. Just because "logon = password authentication" (which is not true, BTW) they believe that "SSO" needs to be password-based, as well. And that passwords need to be kept identical in all systems. That's stupid. SSO is <u>not</u> based on password authentication. Actually, the users do not need to have any passwords anymore (except in the single system where they logon using userID / password, e.g. the Enterprise Portal server).

Noone would attempt to use local Windows user accounts and to synchronize them in the entire network landscape. Noone would attempt to copy the /etc/passwd file in an UNIX server landscape, either. But why attempting the same for SAP systems?!

Well, you could opt for a domain controler approach. That's something I'd fully support ...

Re: Field Size differences in tables when upgrading to Unicode

Former Member — Mon, 21 Aug 2006 11:30:21 GMT

Dear Wolfgang,

I agree on your comments on password synchronization tools. Beeing an auditor I had some "fun" with badly implemented password synchronization tools.

You mentioned that there has been a change to USR02. Can you elaborate on that a bit ?

Thanks and regards,

Ralf

Re: Field Size differences in tables when upgrading to Unicode

Wolfgang_Janzen — Tue, 22 Aug 2006 11:55:21 GMT

Hi Ralf,

well, the structure of table USR02 has changed several times (e.g. between 6.20 and 6.40 respectively when applying the combined ABAP/kernel correction of note 450452 and once again between 6.40 and 7.0). Each time, new features have required that modification.

Once again: the structure of internal data structures or tables might change at any time. Only APIs are stable.

Of course, we take care that such modifications will not cause problems when performing upgrades. After an upgrade all users are still able to logon - that is an absolute requirement. And also we take care that the CUA (Central User Administration) can still be used in all possible scenarios (older/newer CUA master/client systems), sometimes with some functional restrictions, when it is unavoidable.

Regards, Wolfgang

PS: I'm aware that Justin has also posted a customer message regarding this issue (I've been notified by some support colleagues via inhouse mail). According to that customer message only read access (to table USR02) is performed. Yes, Unicode can make a difference regarding offsets and field lengths (of character-wise fields), of course. I'd recommend to query the DDIC (SE11: display "runtime object") to determine this information.