I'm having the same problem as detailed in the forum for SQL Data Generator 2 here.
To recap I have these tables:
dbo.TableA (KeyA int primary key)
dbo.TableB (KeyB int primary key)
dbo.MapA2B (KeyA int, KeyB int, unique(KeyA, KeyB))
I've used SQL Data Generator (126.96.36.199) to generate 90 rows for dbo.TableA & 1,400 rows for dbo.TableB.
The next task I'm trying to do is generate data for dbo.MapA2B. For that I have set:
KeyA - repeat 1 - 2 times
KeyB - repeat 1 - 100 times
I've also set "When data is invalid" to "Skip row".
When I run the data generation I get this error:
Violation of UNIQUE KEY constraint 'XXX'. Cannot insert duplicate key in object 'dbo.MapA2B. The duplicate value is (YY, ZZZ). The statement has been terminated.
I've found some old (4 years, SQL Data Generator 1) references to this being a problem but I can't find any solutions. To be honest I'm rather surprised it has not been fixed. The many-to-many relationship pattern is pretty common.