Am Dienstag, 15. August 2006 00.42 schrieb Abu Mushayeed:
> We upgraded our Db from LATIN1 to UTF8 as follows:
>
> 1. new install with encoding=UTF8
> 2. pg_dump from old DB and import into new db using psql as follows:
> pg_dump out of old_db|psql into new_db
> 3. New feeds (data files) come in same as when the DB was latin1. If error
> message from any load script like "Invalid UTF8 code ...", then, "set
> client_encoding=latin1" in the script and run again and all is well.
So most probably the dump is latin1 (this was the clientencoding during
dumping).
> 4. all
> the other scripts run on the UTF8 Db as is and no errors, meaning the
> scripts has not been changed and is running as is before when DB was
> latin1.
That there is no error message does not mean that there is no error in the
encoding. It only means that there is no Invalid UTF8. So it is importent to
set the correct client encoding, which was used for the backup, also during
restore.
> 5. Now, some of the DB feeds (data files) is changing to UTF8. What
> needs to be done when the data feeds changes to UTF8.
When you plan to insert non Latin1 codes into the database you have to change
the clientencoding for backup/restore to UTF8.
Have a look at chapter 20 of the manual for different possibilies to set the
client encoding.
Best regards
Ivo
>
> Any other thoughts or if there is a problem that anyone sees, please help
> me with a solution.
>
> Thanks
>
>
> ---------------------------------
> Talk is cheap. Use Yahoo! Messenger to make PC-to-Phone calls. Great rates
> starting at 1¢/min.