Hi,

> Feature request: Changes in Unique constraint / Unique Index

You do know that an Unique Index is something totally
different from an Unique Constraint, right?

> From the manual:
>
> The constraint allows existence of only those rows for which the
> aforementioned <search condition>
> evaluates to True. In a unique index or under a UNIQUE constraint, two
> sets of column values will be
> considered distinct and thus allowed if:
> a) both sets contain only nulls, or
> b) there is at least one pair of corresponding values of which one is
> non-null, and the other either null
> or a different non-null value.
> .
> .
> .
>
> What I would like is that two or more sets of column values will be
> considered distinct and thus allowed if: at least one contains nulls.
>
> After all isn't it safe to say that (1,null) is distinct from (1,null) as
> long as null is distinct from null ?

With regards,

Martijn Tonies
Database Workbench - developer tool for InterBase, Firebird, MySQL & MS SQL
Server.
Upscene Productions
http://www.upscene.com