Kevin Falcone wrote:
Temporarily drop the unique constraint on Name and rerun the
conversion, then look for duplicated usernames / usernames that are
blank. That'll give you an id that you can use to clean up the name
in the original content before you do your final cutover.
Thanks Kevin,
I've been doing a test migration from RT 3.6.5 to 3.8.7 on a spare VM,
prior to doing it on our live server. I noticed there were a lot of db
changes, and wanted to be sure that it would all be smooth, and that I
knew how long it would be down. Good thing I did, because while running
the
On Tue, Jan 26, 2010 at 04:53:58PM +, Howard Jones wrote:
I've been doing a test migration from RT 3.6.5 to 3.8.7 on a spare VM,
prior to doing it on our live server. I noticed there were a lot of db
changes, and wanted to be sure that it would all be smooth, and that I
knew how long it
This is due to the UTF8 collation recognizing accents as being the same.
The old latin collation saw them as unique. There's no real fix other
then going utf8_binary but that would make everything case sensitive.
Basically latin1 in mysql is CI_AS and utf8 is CI_AI. For me it was
the email