> I now want to be able to give the option of opening a second database file, > with the same table structure - which i believe is acheived using the ATTACH > command. However the ids sent to the client still need to be unique, it is my > understanding that a SELECT command will treat both databases as one and so > return potentially overlapping ids. Is there anyway to craft the SELECT > command - or otherwise - so that i can know where the id came from and do > some form of id mangling, keeping in mind i will need to do the same process > in reverse to access the information, given the mangled id.
If you attach the second database a join will let you connects rows in one table with rows in another table. Then a select can return 2 columns, the id from each table. The two id's concatenated will create a simple unique key that can be used to find rows in either table. example: two tables, named 'a' and 'b' select a.id, b.id from a inner join b on b.something = a.something result: a.id b.id 23 34 create a single key by using 6 digits (or binary if you understand that) for each: 000023000034 There's your unique key that can be used to find rows in either or both tables.