Skip to content

Conversation

@Aggarwal-Raghav
Copy link
Contributor

@Aggarwal-Raghav Aggarwal-Raghav commented Nov 21, 2025

What changes were proposed in this pull request?

HIVE-29258

Why are the changes needed?

This PR removes explicit Class.forName calls, leveraging the JDBC 4.0+ Service Provider Interface (SPI) for automatic driver discovery. The DriverManager now utilizes ServiceLoader to locate drivers defined in META-INF/services and registers them dynamically.

Does this PR introduce any user-facing change?

NO

How was this patch tested?

Ran the affected files UT locally and will see CI outcome.

Copy link
Contributor

@okumin okumin left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We can amend some points, following SonarQube.

I wonder if we can remove some more. Probably, the answer is no.

% git grep 'Class.forName' | grep Driver
beeline/src/java/org/apache/hive/beeline/BeeLine.java:        Driver driver = (Driver) Class.forName(clazzName, true,
beeline/src/java/org/apache/hive/beeline/Commands.java:        Driver driver = (Driver) Class.forName(name).newInstance();
beeline/src/java/org/apache/hive/beeline/DatabaseConnection.java:            (Driver) Class.forName(clazzName, true, Thread.currentThread().getContextClassLoader())
jdbc-handler/src/main/java/org/apache/hive/storage/jdbc/JdbcStorageHandler.java:      classesToLoad.add(Class.forName("com.mysql.jdbc.Driver"));
jdbc-handler/src/main/java/org/apache/hive/storage/jdbc/JdbcStorageHandler.java:      classesToLoad.add(Class.forName("org.postgresql.Driver"));
jdbc-handler/src/main/java/org/apache/hive/storage/jdbc/JdbcStorageHandler.java:      classesToLoad.add(Class.forName("oracle.jdbc.OracleDriver"));
jdbc-handler/src/main/java/org/apache/hive/storage/jdbc/JdbcStorageHandler.java:      classesToLoad.add(Class.forName("com.microsoft.sqlserver.jdbc.SQLServerDriver"));
jdbc-handler/src/main/java/org/apache/hive/storage/jdbc/JdbcStorageHandler.java:      classesToLoad.add(Class.forName("com.ibm.db2.jcc.DB2Driver"));
ql/src/java/org/apache/hadoop/hive/ql/DriverFactory.java:          Class<?> cls = Class.forName(name);
ql/src/test/org/apache/hadoop/hive/ql/TestDriverFactory.java:          Class<?> cls = Class.forName(name);

If we can reduce the usage, is it possible to remove CONNECTION_DRIVER in HiveConf and MetastoreConf?

public void setUpBefore() throws Exception {
if (miniHS2 == null) {
Class.forName(MiniHS2.getJdbcDriverName());

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change

An extra line has been added to some files and not been added to some files. I prefer to have no extra lines consistently

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Addressed.

}
try {
LOG.info("Going to load JDBC driver {}", driverName);
driver = (Driver) Class.forName(driverName).newInstance();
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This might change the behavior of L67 because this.driver is never assigned now

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Removed driver , connString as class attributes. Made connString local to method getConnection()

@Aggarwal-Raghav
Copy link
Contributor Author

We can amend some points, following SonarQube.

I wonder if we can remove some more. Probably, the answer is no.

% git grep 'Class.forName' | grep Driver
beeline/src/java/org/apache/hive/beeline/BeeLine.java:        Driver driver = (Driver) Class.forName(clazzName, true,
beeline/src/java/org/apache/hive/beeline/Commands.java:        Driver driver = (Driver) Class.forName(name).newInstance();
beeline/src/java/org/apache/hive/beeline/DatabaseConnection.java:            (Driver) Class.forName(clazzName, true, Thread.currentThread().getContextClassLoader())
jdbc-handler/src/main/java/org/apache/hive/storage/jdbc/JdbcStorageHandler.java:      classesToLoad.add(Class.forName("com.mysql.jdbc.Driver"));
jdbc-handler/src/main/java/org/apache/hive/storage/jdbc/JdbcStorageHandler.java:      classesToLoad.add(Class.forName("org.postgresql.Driver"));
jdbc-handler/src/main/java/org/apache/hive/storage/jdbc/JdbcStorageHandler.java:      classesToLoad.add(Class.forName("oracle.jdbc.OracleDriver"));
jdbc-handler/src/main/java/org/apache/hive/storage/jdbc/JdbcStorageHandler.java:      classesToLoad.add(Class.forName("com.microsoft.sqlserver.jdbc.SQLServerDriver"));
jdbc-handler/src/main/java/org/apache/hive/storage/jdbc/JdbcStorageHandler.java:      classesToLoad.add(Class.forName("com.ibm.db2.jcc.DB2Driver"));
ql/src/java/org/apache/hadoop/hive/ql/DriverFactory.java:          Class<?> cls = Class.forName(name);
ql/src/test/org/apache/hadoop/hive/ql/TestDriverFactory.java:          Class<?> cls = Class.forName(name);

If we can reduce the usage, is it possible to remove CONNECTION_DRIVER in HiveConf and MetastoreConf?

Have removed reflection code changes from BeeLine.java, Commands.java, DatabaseConnection.java. Please review it. Changes in JdbcStorageHandler.java, DriverFactory.java are not specific to DriverManager and are outside the scope of PR IMO.

@sonarqubecloud
Copy link

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants