We'd like to prevent record duplication in our MS access database using a multicolumn unique index. Because of how the data is sent (via network), duplicate data is sometimes received. The data source does not send a unique ID, so the simplest option is to prevent duplicate records being inserted.
According to Unique Index Design Guidelines:
With multicolumn unique indexes, the index guarantees that each combination of values in the index key is unique. For example, if a unique index is created on a combination of LastName, FirstName, and MiddleName columns, no two rows in the table could have the same combination of values for these columns.
This is for SQL 2005 however, so I'm not sure it's possible using MS access.
I guess an alternative is to perhaps use the query (pseudo code):
insert into foobar (a, b, c) values ('x', 'y', 'z')
where (a <> 'x') and (b <> 'y') and (c <> 'z')
... but I feel like an index would be better.