Description:
When creating/working with relationships in WB's modelling interface, whether or not a relationship is "identifying" is automatically detected based on whether the key columns are part of the PRIMARY KEY or not, however whether or not the relationship is 1-n or 1-1 is not detected automatically and must be configured manually.
The various relationship-creation tools go some way to handling this (creating the appropriate type of relationship) but if a relationship is to be altered then it must be edited manually by double-clicking the relationship line and editing it's foreign key setting.
Actually editing this setting (toggling between 1-to-many and 1-to-1) seems to have no effect except updating the rendered line in the diagram. I would think that "the act of making a relationship one-to-one" in a diagram would also take whatever measures are required to enforce such a relationship in the schema (that is, the creation or removal of a UNIQUE INDEX on the foreign key column). Also the creation of indexes that would result in forcing a relationship to become 1-to-1 could also immediately change the rendered state of a relationship line accordingly.
Would it be useful to link this functionality with the creation of a UNIQUE INDEX (or removal of that index) on the foreign key column(s)? This index-creation would actually enforce that 1-to-many or 1-to-1 relationship status within the schema.
Equally, in a similar way to how a relationship "instantly becomes" an identifying relationship if it spots that its foreign key columns are added to that table's primary key, a relationship could "instantly become" 1-to-1 if a unique index is added covering the relationship's foreign key columns (and vice-versa becoming 1-to-many if such an index is removed).
How to repeat:
N/A - feature request
