This content has been marked as final. Show 3 replies
go to document 1086400.1 as specified in that.
and in that document see the note
"If a "select * from sys.sys_tzuv2_temptab;" give(s) affected columns then you need to back up the reported columns before upgrading to 10.2.0.5. See point 5) on how to do this. For SYS.SCHEDULER$% rows you need to drop and re-submit the jobs (see previous point)."
You get that by running the script supplied by oracle which really all that is doing is going off dba_tab_columns to see what tables have time zone aware columns.
As further explained the purpose of this is they want you save off the timezone aware data so that after you upgrade your engine to 10.2.0.5 and the resultant timezone files are upgraded you can update the data back (or re-import or whatever) so that oracle is using the version 4 timezone files as the data goes back in.
My question still stands though. The only difference between 3 and 4 are some daylight savings changes I think.
So for example a piece of data that has timezone info in it, but is in, say, July is completely uneffected by this.
So I'm getting this info out of the same document # that you are listing... Granted oracle's documentation is rather byzantine to wind your way through but if you follow the breadcrumbs around that's basically it in a nutshell.