Let's say I have a table (I, do they are very cheap these days) The table has a composit primary key:
create table ( a int not null, b int not null, c int not null, ... more columns here. primary key(a,b) ) Now, it just so happens that a is a foreignkey that I'd like to look up rows by quite often, so the natural thing to do here is to create an index on a... However, I have this idea; that because a is the first column of an index already, that this index will be just as fast as an index just for a, but is it? What do I need to think about before re-using an index like this? You don't need to maintain an extra index. If the other index is very large it could be a loss to wade through it. If the other index is used a lot then it would more likely be in cache. At the moment I don't generate indexes for columns that are already indexed as the first column of another index (primary key or one created manually), I think this is a winning strategy, but what do other people think? -- Regards Flemming Frandsen aka. Dion/Swamp http://dion.swamp.dk _______________________________________________ sapdb.general mailing list [EMAIL PROTECTED] http://listserv.sap.com/mailman/listinfo/sapdb.general
