support for non-latin characters

mbojanks · Joined: 06 Aug 2008 Posts: 10

Please tell me if there is a support for non-latin character transformation, particularly from MS Access and MySQL. When I tested MS Access database with Unicode chars to transform to other MS Access database and to MySQL, question marks appear instead of non-latin chars. Thanks

Sergei · Joined: 26 Jul 2007 Posts: 969

mbojanks · Joined: 06 Aug 2008 Posts: 10

Those question marks appear in the Preview results window, both from target MS Access and MySQL connectors... I tested on my local windows xp machine with JRE 1.6.0_07, both times I tried to transform (without any function in between) Serbian Cyrillic unicode data from one Access table into another (the first test) and then into a MySQL database table (the second test), and both times I obtained question marks in the target table's connector Preview results window... Perhaps batch file should contain encoding switch for proper operation on non-latin chars (to compile into different encoding)... Thanks for your quick response

alex · Apatar Team Member Joined: 30 Mar 2007 Posts: 464

mbojanks · Joined: 06 Aug 2008 Posts: 10

Thank you Alex for your support and for regards as well. I don't quite understand what sort of table you tested on (plain text or something?) when you had Cyrillic characters previewed correctly. What about MySQL tables? I also have the question mark issue there. I still hope it will be solved soon, because the Apatar project deserves ultimate success in its idea of connecting and transforming data from anywhere to everywhere. I 'm sure Apatar developers will manage to find the solution.
Kindest regards,

alex · Apatar Team Member Joined: 30 Mar 2007 Posts: 464

mbojanks · Joined: 06 Aug 2008 Posts: 10

Hello, Alex. I hope the main developer will start his/her new working season with success in solving this bug. Perhaps s/he should analize MySQL connector as well. I wish you all the best to all of you in Apatar.
Regards,
Bojan

Michael · Joined: 26 Aug 2008 Posts: 22

mbojanks · Joined: 06 Aug 2008 Posts: 10

When I tested, both databases (source and target) were on the same machine and the question mark issue occurred. Those were both .mdb files from my local computer. If I can do anything to provide you with more information, be free to ask me for.
Thank you for your effort in solving this problem. I am sure you will manage to solve it and make Apatar even more versatile tool...
Speaking of versatility, what about support for EDI FACT format that some commercial products provide (Softshare Delta for instance)?

mbojanks · Joined: 06 Aug 2008 Posts: 10

OK, now, I see your point. I 'll try to set Serbian Cyrillic encoding on my computer with Apatar window in focus and then try the testing transformation again.
Thank you, Michael, for providing this solution.

mbojanks · Joined: 06 Aug 2008 Posts: 10

... and here is the result:
I clicked on my bottom-right corner SR button while Apatar window in focus and make Serbian Cyrillic active keyboard, then I clicked on Preview results for output MS Access connector (source table has 840 records of places - MESTO=PLACE; target table is also MESTO with no records) and obtained this in my Preview window:
Progress:
com.altoros.octoslave.msaccess.MsAccessNode created.
Records Processed 0
Start Reading from MESTO
Records Processed 840
com.altoros.octoslave.msaccess.MsAccessNode created.
Records Processed 0
Start Reading from MESTO
Records Processed 840
com.altoros.octoslave.transform.TransformNode created.
Records Processed 0
com.altoros.octoslave.msaccess.MsAccessNode created.
Records Processed 0
Start writing to MESTO
Insert Mode
Records Processed 676
[Microsoft][ODBC Microsoft Access Driver] Cannot open any more tables.
Start Reading from MESTO
Records Processed 3380
Finish
//end of Preview window data

Results previewed: 2704 records with ???? instead of Cyrillic chars
2704=4*676→ maybe it means something
There is also the following issue: Nothing happens in the target MESTO table (remains empty), when I click Run and the same Progress: is written
You would know better this ODBC Driver error is all about, it could be the cause of not performing the transformation when Run was clicked... and why proccessing all this for four times
The 677th record is nothing special (Томашевац - if you can see it in Cyrillic)
Also, my JRE Version is 1.6.0_07, as displayed in DEbug Console, but I also have 1.5.0_15 installed

I hope these details can help you locate a problem.
Regards,
Bojan

mbojanks · Joined: 06 Aug 2008 Posts: 10

Naturally, MESTO source and target tables are in different .mdb files...

Michael · Joined: 26 Aug 2008 Posts: 22

Hi Bojan,

Are you using an English or Serbian version of Windows? If you have an English Windows version, try installing Serbian character encodings via Control Panel>Regional and Language Options>Advanced.
_________________
Michael Fedotov
Apatar Evangelist
Contact Apatar Team

mbojanks · Joined: 06 Aug 2008 Posts: 10

I use English Windows. I had problems with loading the operating system when I tried Language for non-Unicode programs to be Serbian (Cyrillic). I have Serbian language installed for keyboard layout and all Cyrillic Code page conversion tables checked.
I performed another test: included just five records in the source table and executed Preview results on INPUT CONNECTOR. I concluded from the results that DATA ARE INITIALLY NOT SEEN PROPERLY, even with Serbian Latin chars (for instance, š is seen as љ, Đ as P, Ć as C, and all Cyrillic chars as question marks). Transformed chars are not seemed to be a problem, EXCEPT FOR RUN NOT WORKING - output table remains empty!
I really wish you to solve these problems, because it would bring a very good and useful software to entire community.
Regards,
Bojan

rtest · Joined: 16 Dec 2008 Posts: 71

I am also facing the same problem with "Thai" language characters.

Any code change done in resolution of this problem.

Regards,
Rohit