Using com.databricks:databricks-jdbc:2.6.36 inside oracle stored proc

[Ujeen]

Hi dear Databricks community,

We tried to use databricks-jdbc inside oracle store procedure to load something from hive. However Oracle marked databricks-jdbc invalid because some classes (for example  com.databricks.client.jdbc42.internal.io.netty.util.internal.logging.Log4JLoggerFactory) are importing non-existent package: 

import com.databricks.client.jdbc42.internal.apache.log4j.Logger;

if this was done intentionally inside databricks jdbc lib is it possible to make oracle load that jar still anyhow ?

Thank you 

Sincerely Yours, Eugene

[Kaniz]

Hi @Ujeen , When integrating Databricks JDBC with an Oracle-stored procedure to load data from Hive, encountering issues related to missing classes can be frustrating.

Let’s explore some potential solutions:

1. Check Dependencies:

   • Ensure that all necessary dependencies are correctly configured. Verify that the Databricks JDBC driver and any other required libraries are accessible to Oracle.
   • Confirm that the classpath includes the relevant JAR files.

2. Oracle Invalid Objects:

   • When Oracle marks a package or class as invalid, it means that the referenced objects are not available or are in an inconsistent state.
   • To address this, you can try recompiling the invalid objects using the following SQL command:

     ALTER PACKAGE <package_name> COMPILE;
     

   • Replace <package_name> with the actual name of the package containing the invalid class.

3. Custom Class Loading:

   • Oracle allows custom class loading using the loadjava utility.
   • You can load the required JAR file into the Oracle database using the following command:

     loadjava -user <username>/<password>@<database> -resolve -verbose <path_to_jar_file>
     

     • Replace <username>, <password>, <database>, and <path_to_jar_file> with appropriate values.
   • After loading the JAR, you can create a Java source in Oracle that references the classes from the JAR.

4. Grant Permissions:

   • Ensure that the Oracle user executing the stored procedure has the necessary permissions to access the JAR and its classes.
   • Grant the required privileges using:

     GRANT EXECUTE ON <package_name> TO <user>;
     

5. Debugging:

   • Check the Oracle logs for detailed error messages related to the missing classes.
   • Investigate whether there are any classpath issues or conflicts.

For more specific guidance, consult the relevant documentation for both Oracle and Databricks¹².

Hopefully, you’ll soon have your Databricks JDBC integration working seamlessly! 🚀