Based On Python, Mysql Database Application Layer Level Partitioning Technology

In the modern information processing, the ever-growing amount of data stored in a single database is difficult to meet the needs of the database partition level growing demand. At present stage, MySQL database has many restrictions on partition, such as the partition will not support foreign key references, non-integer arithmetic operations and bitwise operations, the primary key table type must be an integer in partition, the number of partition is limited, and the partition primary key cannot be a sub query and so on.As many products that use the MySQL database, in order to not make any changes or make the fewest changes to the database under horizontal partitioning, and to resolve the MySQL database partition restrictions, so that the original investment will be preserved as far as possible.Thus, we need a middle layer between applications and MySQL database, so that we can resolve the above limitations, while retaining the original SQL statement syntax and application interfaces to improve the MySQL database restrictions on the horizontal partition, and also will be able to make the original product needn't make any changes to adapt to the partition of MySQL database, so making the original investment will be preserved as far as possible.The Python scripting language is simple and more suitable for describing business logic, the speed is also in the forefront of many scripting languages and many products are developed by the use of the Python language.This thesis based on the Python language and starts from the analysis of SQL statements, then split the SQL statement in the multi-table query into multiple single-table queries and multiple databases from multiple single-table query results summarized under the joint conditions, and return the original SQL execution results before partition. And use the real load test shown that in a single table query result set is not a great situation can get close to the linear expansion ratio.