SQL capabilities

Transaction

Complete support for transaction ACID features, MVCC and  committing two isolation levels: read and repeatable read

Built-in functions

Full support for MySQL built-in functions

User Defined Function

Support for user defined functions

View

Support for views

Stored procedure

Support for stored procedures

Trigger

Support for triggers

DQL statement

Support for SQL2016 standard

DDL statement

Support SQL2016 standard, default index type is Hash index. Btree index needs to be specified by using btree keyword

DCL statement

Support for SQL2016 standard

TCL statement

Support for SQL2016 standard

Distributed capabilities

Distributed transactions

Complete support for distributed transactions, which are completely transparent to the application

Global table

Each node has the full amount of data in the table, which is suitable for small tables with more reads and fewer writes, and improves the efficiency of correlation queries

Hierarchical table

Tables have hierarchical relationship with each other. The primary key of child table needs to contain the partition key of the parent table, that is, the child table is distributed according to the distribution key of the parent table by default.

Distributed table

The data in the table is distributed to different nodes according to the policy, and the distribution key can be either a Hash index or an interval index

Hash slice

When the distribution key of the distribution table is a Hash index, the data is partitioned by Hash. Data is automatically balanced across nodes.

Range slice

When the distribution key of the distribution table is a BTree index, the data is sliced by Range to ensure that the consecutive data are in the same node as much as possible. The data is automatically balanced across nodes.

Global secondary index

The global secondary index is the mapping between the index attribute and the shard index. Data retrieved through the global secondary index will be directly mapped to the shard index, thus avoiding broadcasting.

Operational migration

Custom functions can specify routes to migrate to the appropriate node for execution

Online scaling

Support for non-stop service and online scaling of nodes. Data can be automatically balanced

Load balancing

Load balancing clients connect to different nodes

Generic Capabilities

Access control granularity

Provides row-level access granularity control through table locks and row locks

Multiple development interfaces

Support for C, C++, API, ODBC, JDBC, Python, and other database development interfaces

Character set support

Support for utf8, gbk, latin1 and other character sets

Security

Users, Roles, Permissions, Auditing

Support user management, role management, access rights control, and audit log

Disaster tolerance and recovery

High availability

Strong financial-level consistency via Raft protocol, with automatic master-slave switching

Data backup

Support data snapshot backup, full-snapshot write without blocking read and write operations of database

Data recovery

Support fast recovery in global and single-node dimension by backup data