Upgrade of Devlopment box using CU&UC is done. How to verifiy?

[Former Member]

Dear team,

We need your Advise at earliest possible as requested by our

Management.We have performed CUUC ( Combined Upgrade & Unicode

Conversion ) of our

Development System from 4.6C to ECC 6.0 on HP-UX 11i v3.

We are a bit concerned about the Unicode Conversion steps performed. Since we are not sure of the language assignments during upgrade if they were correct or not.

We opened a message with SAP and the reply was this is not a support issue but rather consulting.

How can we check/verify our system before proceeding to QA and production ?

Regards,,,

[markus_doehr2 · Accepted solution]

Which languages did you use productively before you migrated to Unicode and which languages do you have installed?

--

Markus

[Former Member · Accepted solution]

Hi,

Sorry that I did not mention this.

Our system is MDMP with 8 languages but only English was used. The other languages were future plans. (Arabic - Chinese - English - French - German - Italian - Japanese - Spansih).

[markus_doehr2 · Accepted solution]

I see.. so you have several codepages...

English

French

German

Italian

Spanish

are all the same (ISO8859-1 resp. 8859-15)

What comes into place now are the other languages.

I'd login in all of those and check, if the text is correctly displayed.

How did you process your vocabulary? DId you logon with each language with native language speakers or did you do this "yourself" by guessing? Just curious, because we migrated last year an MDMP system with 12 languages...

--

Markus

[Former Member · Accepted solution]

Hello,

Really thanks for your quick reply.

To tell you the truth. We had a partner onboard and did the whole thing without native speakers !! So to answer you it is guessing!

That's why I posted the question. So where/how can I find an answer ?

[markus_doehr2 · Accepted solution]

It's possible (we did it likewise - well, almost).

If you logon in Arabic or Chinese (someone, who can basically read and understand that language), do they see the transaction texts correctly?

E. g. if you login in Chinese, do you see any weird characters (like you see if you logged on to the MDMP system in english and checking chinese vocabulary)?

--

Markus

[Former Member · Accepted solution]

hello,

I did what you mentioned.

The screens I saw were O.K. but : (and here's the catch!)

SE11 - > Table: ZARABEMPNAME 5 entries in the 4.6C system correctly showed arabic charecters (readable & correct).

the same procedure in the the ECC system gave me garbage characters. Menaing I have corrupted data.

(both systems exist together since 4.6 system exists on AIX and we did an HP Smooth transition method to another box on HP-UX then upgraded.)

I think this evidence that the is a corruption... Right ?

[markus_doehr2 · Accepted solution]

Hmmm..

which logon language did you use on the 4.6c system to show the data in SE11? If it was english, then you violate an MDMP law: In an MDMP system only the data of the login codepage must be entered. If someone entered arabic using an english logon, then it's clear, that the data is corrupt.

--

Markus

[Former Member · Accepted solution]

Thanks,

The old system refuses to log me in Arabic. The Arabic characters I saw were in fact in Arabic but i have no idea how they were entered.

The new system when logged in in Arabic shows garbage characters!!

[markus_doehr2 · Accepted solution]

So yes - then this error is expected.

Someone entered that data apparently using an English login - and this is WRONG, one MUST NOT do that because the data is stored in the database with the logon language, thus the data is in ENGLISH ASCII characters.

If you do the same thing again (logon in English and display), then it's correctly displayed.

The Unicode conversion took the language key (which is is EN since the data was entered with an english logon) and converted it accordingly. Now it's corrupted due to this.

There's nothing you can do but correct it manually because the user did the mistake of entering that data with the wrong logon language.

--

Markus