ok, i am assuming that you want a name for all id's. if a has an entry use that, otherwise use the name from b. also assuming that B has an entry for every id. if those assumptions are correct, then this should work

Or maybe:

that would add the assumption that the names in b are the first letter of the names in a

Hmm not really, it replaces the assumption that a.id is a foreign key to b.id.

The OP does not specify. If there is a relation between the id's (which there should be in my opinion, and if the data provided represents all possibilities, then this is indeed the case) your solution is the more efficient.

However if the id's are not related then that only leaves the initial letters, in which case my alternate solution will do.

Thanks to reply my query. It worked.