A, JDBC driver
JDBC driver is loaded:
1, by loading the container:
Java applications for the container, the corresponding drive can be directly placed in a container jar package lib directory, for example in a container made of Tomcat web application, the driving copied to the tomcat lib subdirectory;
2, load the application is running:
If you want to apply self-loading the appropriate drivers, you need to specify a good search maven classpath jar when the drive when playing for the application jar / war package, please refer to:
https://maven.apache.org/shared/maven-archiver/examples/classpath.html
https://www.cnblogs.com/snaildev/p/8341610.html
(With the jar package lib directory store executable dependent jar package) Example 1:
<build> <finalName>image</finalName> <plugins> <plugin> <groupId>org.apache.maven.plugins</groupId> <artifactId>maven-jar-plugin</artifactId> <version>2.4</version> <configuration> <archive> <manifest> <addClasspath>true</addClasspath> <classpathPrefix>lib/</classpathPrefix> <mainClass>some.package.MainClass</mainClass> </manifest> </archive> </configuration> </plugin> </plugins> </build>
When the traffic load JDBC driver itself, the following operations need to be performed prior to loading the application using JDBC Driver:
Class.forName("com.mysql.jdbc.Driver");
// Then the following begin to use jdbc
Second, (a kind DataSource) is connected via the database connection pool manager
Currently used for the connection pool or Druid Hikari, respectively, as follows:
1, Druid connection pool:
maven dependence:
<dependency> <groupId>com.alibaba</groupId> <artifactId>druid</artifactId> <version>1.1.11</version> </dependency>
Sample code:
DruidDataSource createSource(String name, String jdbcUrl, String userName, String password,
maxActive, minIdle, maxWait, scanInterval, minActiveTime) {
DruidDataSource src = DruidDataSource();
src.setName(name);
src.setUrl(jdbcUrl);
src.setUsername(userName);
src.setPassword(password);
src.setDriverClassName();
src.setInitialSize(minIdle);
src.setMaxActive(maxActive);
src.setMinIdle(minIdle);
src.setMaxWait(maxWait);
src.setTimeBetweenEvictionRunsMillis(scanInterval);
src.setMinEvictableIdleTimeMillis(minActiveTime);
src.setTestWhileIdle();
src.setTestOnBorrow();
src.setTestOnReturn();
src.setPoolPreparedStatements();
src.setMaxPoolPreparedStatementPerConnectionSize();
src.setValidationQuery();
src.setRemoveAbandoned();
src.setRemoveAbandonedTimeout();
src.setKeepAlive();
if(!isDataSourceOk(src)){
LOGGER.error("Data source "+ name + "test failed");
}
return src;
}
private static boolean isDataSourceOk(DataSource source){
try (Connection connection = source.getConnection();
PreparedStatement stmt = connection.prepareStatement("select 1");
ResultSet resultSet = stmt.executeQuery()){
resultSet.next();
return true;
}catch (Throwable e){
return false;
}
}
2, Hikari 连接 pond
maven dependence:
<dependency> <groupId>com.zaxxer</groupId> <artifactId>HikariCP</artifactId> <version>3.2.0</version> </dependency>
Sample code:
HikariDataSource createSource(String name, String jdbcUrl, String userName, String password,
connectionTimeout, idleTimeout, maxLifeTime, poolSize) {
HikariDataSource src = HikariDataSource();
src.setPoolName(name);
src.setJdbcUrl(jdbcUrl);
src.setUsername(userName);
src.setPassword(password);
src.setDriverClassName();
src.setConnectionTimeout(connectionTimeout);
src.setIdleTimeout(idleTimeout);
src.setMaxLifetime(maxLifeTime);
src.setMaximumPoolSize(poolSize);
if(!isDataSourceOk(src)){
LOGGER.error("Data source "+ name + "test failed");
}
return src;
}
三、处理连接
1、首先通过连接池获取连接，例如：
try(Connection getConnection() SQLException {
assert DataSourceManager.getDataSource() != null;
return DataSourceManager.getDataSource().getConnection();
}
2、连接的SQL请求提交方式
从连接池（无轮DruidDataSource还是HirakiDataSource）申请到Connection对象拿到后，默认的提交方式为自动提交，即此时调用connection.getAutoCommit()返回的一定是true；
Connection对象归还连接池后，下次再从连接池申请一个Connection，默认的提交方式还是自动提交。需要清楚什时候用自动提交、什么时候适合手工提交。
概况起来就是，看申请到Connection对象到归还Connection对象这期间使用同一个Connection对象做了什么操作：
（1）只有读操作，没有任何写操作：用自动提交；
（2）只有一次对单个表的单条记录有进行写操作：用自动提交；
自动提交的代码示例如下
try(Connection connection = ()) {
block.apply(connection); // Do all query or only one update for only one record
}
（3）对同一张表的多条记录进行了写操作，或者对不同表的记录分别进行了写操作：根据是否需要回滚、性能要求，确定是否需要支持事务性；如果要支持事物性，必须采用手动提交；
手动提交的操作示例：
try(Connection connection = ()) {
boolean success = false;
try{
T t = block.apply(connection); // Use this connection process one transaction
doCommit(connection);
success = true;
return t;
}finally{
if(!success) {
doRollback(connection); // If possible, support rollback when failed
}
}
}
上述复杂过程的block中可能对connection的提交模式进行了修改，为了保持代码的兼容性，上述doCommit()及doRollback()的设计如下：
public static void doCommit(Connection connection) throws SQLException {
if (!connection.getAutoCommit()) {
connection.commit();
}
}
private static void doRollback(Connection connection) throws SQLException {
if (SUPPORT_ROLLBACK && !connection.getAutoCommit()) {
connection.rollback();
}
}
（4）耗时的操作，数据量比较大，这时依赖数据库的事物性及回滚已经没法达到；这种情况下应该分多次提交，并有应用层提供回滚；
示例如下：
try(Connection connection = ()) {
boolean success = false;
try{
block1.accept(connection);
doCommit(connection);
block2.accept(connection);
doCommit(connection);
success = true;
}finally{
if(!success) {
block2.clear(connection);
block1.clear(connection);
}
}
}