Effective 1 to n linking

[Simon Remppis]

I have a table A with fairly large records. Each of those records is related to zero or more records of another table B. I always wondered about the most efficient way to read a record from A and all related records from B out of the Database.
Of course I can use a left join, but then mysql returns n rows, all of which contain the large amount of data stored in the single record from table A.
To get rid of this redundant data, I tried to group the result by the primary key of table A and use GROUP_CONCAT to still retrieve all of the B-records. The query then looks like this:
SELECT A.*, GROUP_CONCAT(B.field1) AS field1, GROUP_CONCAT(B.field2) AS field FROM A LEFT JOIN B ON B.aid = A.id GROUP BY A.id
This works, but has some limitations: I have to split the concatenated B-records on the client side and therefore need a seperator, that must never occur in any of the B-records. The maximum length of GROUP_CONCAT is also limited (I know that you can set this limit). Generally, this doesn't look like a very "clean" solution to me.
The third approach that came to my mind is to just use two queries, one for A and one for B.

Are there better solutions for this quite common problem? Which one would you recommend?