2C.Separation and Sharding

Add a state to the inflated data, stop the time, divide and conquer, and combine.

Use ShardingJdbc to handle RW separation, data sharding to handle large data smoothly. Wings uses shard to broadly represent DataSharding, ReadWrite Splitting/Separation.

2C.1.Shortcoming of Sharding

Because ShardingJdbc has some SQL parsing problems (index,trigger) when executing SCHEMA changes. So when doing SCHEMA and journal operation, use plain datasource and finish it with flywave. Reference SQL parsing engine

It may not be necessary to consider shard tables at the beginning of the business, when needed, you can use flywave to generate tables and migrate data.

By default, DDL, DCL use plain datasource, DML etc. use shard datasource to execute. You can also specify the datasource manually to accomplish customized data operations.

2C.2.Things in SQL Execution

DDL - Data Definition Statements

• ALTER TABLE
• CREATE INDEX
• CREATE TABLE
• CREATE TRIGGER
• DROP INDEX
• DROP TABLE can only be one table at a time
• DROP TRIGGER specify plain table manually, according to shard table naming rules to execute
• TRUNCATE TABLE

Among them, except for the system-generated trigger for journal, manual deletion will cause data inconsistency.

DML - Data Manipulation Statements

• DELETE
• INSERT
• REPLACE
• UPDATE

For SQL other than the above types, please use a comment to force to specify datasource and plain table to skip the SQL auto parsing. Theoretically, you should not use flywave to execute SQL other than DDL,DCL and DML, which is outside the scope of versioning.