The AggregatedDialect only implements canHandle, getCatalystType, getJDBCType. It doesn't implement other methods in JdbcDialect.
So if multiple Dialects are registered with the same driver, the implementation of these methods will not be taken and the default implementation in JdbcDialect will be used.
will throw an exception.
Though the quoteIdentifier is correctly implemented in Spark's MySQLDialect and our AnotherMySQLDialect.