RTB 11.5.2 rtb db dump & load considerations

Posted by cverbiest on 14-Apr-2016 03:58

I recently noticed our rtb repository database is still using codepage iso8859-1, 4KB blocksize and type I tables in the schema area.

I'd like to dump & load the db to

  • Type II
  • 8KB blocksize
  • utf-8 codepage

Any considerations / recommendations

Posted by asthomas on 14-Apr-2016 04:13

Hi Carl,

The latest versions of Roundtable actually come with a rtb.st file that has a suggested Type II storage area structure for a RTB database. I would recommend using this as a starting point for your updated repository.

It has been a while since I last tried to convert a database from ISO8859-1 to UTF-8, and as a recall, I had some issues last time I tried ;-)

With regards to a database like the RTB repository, you have to take into consideration that there may be data in your database that may cause some challenges with a code page conversion. I have seen some issues with special characters in code stored in the database that gets "confused" when it is converted.

A common example of this is code where developers have pasted in things like copyright symbols into the code. This may look ok in e.g. ISO8859-1, and when you change to UTF-8, the characters are translated into something that the compiler is unable to understand.

If you know that you have stuff like this in your code, then you should check that after a conversion. If not, then there is no need to worry ;-)

A good way to check this would be to do the dump and load conversion, and then use the Populate Workspace feature in RTB to extract all the code for a workspace - and then compile and run it to make sure things are as they should be.

HTH

All Replies

Posted by asthomas on 14-Apr-2016 04:13

Hi Carl,

The latest versions of Roundtable actually come with a rtb.st file that has a suggested Type II storage area structure for a RTB database. I would recommend using this as a starting point for your updated repository.

It has been a while since I last tried to convert a database from ISO8859-1 to UTF-8, and as a recall, I had some issues last time I tried ;-)

With regards to a database like the RTB repository, you have to take into consideration that there may be data in your database that may cause some challenges with a code page conversion. I have seen some issues with special characters in code stored in the database that gets "confused" when it is converted.

A common example of this is code where developers have pasted in things like copyright symbols into the code. This may look ok in e.g. ISO8859-1, and when you change to UTF-8, the characters are translated into something that the compiler is unable to understand.

If you know that you have stuff like this in your code, then you should check that after a conversion. If not, then there is no need to worry ;-)

A good way to check this would be to do the dump and load conversion, and then use the Populate Workspace feature in RTB to extract all the code for a workspace - and then compile and run it to make sure things are as they should be.

HTH

Posted by cverbiest on 14-Apr-2016 05:03

Hi Thomas,

thanks.

I'm not going to try to convert using codepage convert. I'm going to do a dump & load

iso8859-1 db + cpsteam utf-8 => utf-8 dump files => utf-8 db + cpstream utf-8

The populate is a good idea for a check

Is there a full rtb.df with correct areas available  or should I just run the installer and create a new db ?

Posted by asthomas on 14-Apr-2016 05:06

Create a new db with Setup and Update and rump this - that's the easiest wqay to get a full .df file.

This thread is closed