The main idea behind this database thingie is that you do not need to fetch the whole dataset to client. If you do need it, which is a rare case, it is a bad system design. In most cases, however, you need only a small subset of stored data, and your job is to select it properly.
Second, the mysql server itself can operate too slow in some circumstances -- e.g. no index on 'UserName' field can produce inappropriate time consumption. How many time it takes when you select above dataset from mysql client itself?
Third, you need a space, a memory to store this dataset locally on your client; did you count how much memory do you need to store your 34748806 rows?
