|
||||||||
|
This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira |
||||||||
|
||||||||
|
This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira |
||||||||
There's 3 different areas that need updated:
1.) "not null unique" syntax on a column.
2.) Creating unique constraints.
3.) Creating primary key constraints.
If #1 occurs, #2 and #3 need to be prevented (and vice versa). Currently, simultaneous #1 and #3 occurrences are not handled at all. For instance, the mentioned test case fails on Oracle since the column is "not null unique" and a primary key constraint is attempted as well.
Alternatively, I believe #1 and #2 are currently handled "properly" (although in a confusing way). If the dialect supports "not null unique", that's used and the unique constraint is prevented. We'd just need to prevent #1 and #3 from occurring at the same time.
One decision needs to be whether or not we use "unique" on the column period. If we continue to do so, handling all 3 conditions needs to be improved. Otherwise, simply prevent "unique" from happening on the column and all and always use constraints. Steve Ebersole, what do you think?