Skip to content

InstallingJDBCDriver.md

Latest commit

 

History

History
91 lines (65 loc) · 4.42 KB

InstallingJDBCDriver.md

File metadata and controls

91 lines (65 loc) · 4.42 KB
Raw

Installing the Databricks JDBC driver

Using MATLAB® it is possible to connect to a variety of databases. The MathWorks Database Toolbox™ is often used to connect to databases in a user friendly manner using ODBC/JDBC connections. The following describes how to connect to Databricks® using Database Toolbox and the Databricks JDBC drivers.

The package supports both Databricks JDBC drivers:

  • Versions >= 2.7.3 and < 3.0.0 (the non-OSS Simba based driver).
  • Versions 3.0.3 and newer (the OSS driver).

The JDBC driver must be downloaded from Databricks. The jarFilePath name-value pair can be used when creating a connection to specify the path to the driver.

Links to the licenses for the Databricks drivers:

Downloading the Driver

The Databricks JDBC driver can be downloaded from: https://www.databricks.com/spark/jdbc-drivers-archive

The correct driver version depends on the Java version used by MATLAB:

Java version Driver version required
Java 8 2.X
Java 11 2.X or 3.X
Java 17+ 3.X

MATLAB's Java version can be checked using the version -java command.

Once downloaded, provide the path to the driver JAR file using the jarFilePath argument:

j = databricks.JDBCConnection(jarFilePath="/path/to/databricks-jdbc-2.8.0.jar");

Driver Selection

When both a jarFilePath and a useDriverType argument are provided, both are used. If only one is provided, the driver type is inferred from the JAR file or the argument.

Driver selected Driver argument Jar file Argument
Argument based OSS/Simba OSS/Simba
OSS OSS Unset
Simba Simba Unset
OSS Unset OSS
Simba Unset Simba

When a JDBC Connection is made using databricks.JDBCConnection the chosen driver is added to the dynamic Java class path prior to the invocation of the driver.

Java version configuration

The OSS driver requires that MATLAB uses a Java environment of v11 or greater, and so is not compatible with MATLAB's included default Java environment. MATLAB's Java version can be changed using the jenv command. If the common JAVA_HOME environment variable is defined the jenv(getenv("JAVA_HOME")) command can be convenient. Alternatively a specific path can be set e.g.: jenv("/usr/local/jre"). The version can be reset using jenv("factory"). MATLAB must be restarted for a change to take effect.

For information on compatible versions see: https://www.mathworks.com/support/requirements/openjdk.html.

MATLAB R2022b cannot use a Java version > 8 and so must use the Databricks Simba driver.

If Java 16 is used the following configuration argument must be passed to Java --add-opens=java.base/java.nio=org.apache.arrow.memory.core ALL-UNNAMED. See: https://github.com/databricks/databricks-jdbc. To configure MATLAB Java startup options see: https://www.mathworks.com/help/matlab/matlab_env/java-opts-file.html. It is recommended to avoid version 16 for this reason and that it is not formally qualified with MATLAB.

Usage

The following documents detail how the various database centric interfaces can be used: