Question Convert database to utf8 takes very long

[Sebastian Isenrath]

I'm going to convert database to UTF-8 using PROUTIL.

We used the following command to start the conversion:

proutil database.db -C convchar convert UTF-8

Before that, I used PROUTIL CONVCHAR CHARSCAN to scan through the existing database to check any

The database is about 60 GB an the conversion is now running for over 36 hours.

Is there a possibility to speed up this process?

Many many thanks for any help.

 

[cj_brandt]

We have been better off dumping the database and creating a new db from scratch as UTF-8 and then loading the data.

besides the 36+ hours for the utility to run you still need the index rebuilds.

 

[RealHeavyDude]

proutil convchar charscan will scan each field in each record of type character. Depending on your database and the speed of your disks that can be a lot. Since you don't tell anything else about your environment, although 36 hours seem a lot of time, there is not much to say here.

It's been a long time since I converted a database to UTF-8. What is the current code page the database uses? Usually you would run the charscan either when you convert to target code page that is not compatible or you are not sure whether there are characters in the database that are not part of the character set of the source code page. If the latter is not the case and you are converting to UTF-8 you should not need to do the charscan since converting to UTF-8 should be possible from all code pages that Progress supports.


Heavy Regards, RealHeavyDude.