Apache Spark Connector for Microsoft Fabric

Apache Spark Connector lets you connect to Apache Spark, a unified engine for large-scale data analytics.

In this article you will learn how to quickly and efficiently integrate Apache Spark data in Microsoft Fabric without coding. We will use high-performance Apache Spark Connector to easily connect to Apache Spark and then access the data inside Microsoft Fabric.

Let's follow the steps below to see how we can accomplish that!

Download Documentation

Apache Spark Connector for Microsoft Fabric is based on ZappySys JDBC Bridge Driver which is part of ODBC PowerPack. It is a collection of high-performance ODBC drivers that enable you to integrate data in SQL Server, SSIS, a programming language, or any other ODBC-compatible application. ODBC PowerPack supports various file formats, sources and destinations, including REST/SOAP API, SFTP/FTP, storage services, and plain files, to mention a few.

Prerequisites

Before we begin, make sure you meet the following prerequisite: Java Runtime Environment (JRE) or Java Development Kit (JDK) must be installed on your system.

Minimum required version: Java 8
- Amazon Corretto 8
- Oracle JRE 8
Recommended Java version: Java 21
- Amazon Corretto 21
- Oracle JRE 21

If your JDBC Driver targets a different Java version (e.g., 11 / 17 / 21), install the corresponding or newer Java version.

If you already have a JRE installed, you can try using it too. However, if you experience any issues, we recommend using one of the distributions mentioned above (you can install an additional JRE next to the existing one; just don't forget to configure the default Java in the Windows Environment Variables).

Download Apache Spark JDBC driver

To connect to Apache Spark in , you will have to download JDBC driver for it, which we will use in later steps. Let's perform these little steps right away:

Visit MVN Repository.
Download the JDBC driver, and save it locally, e.g. to D:\Drivers\JDBC\hive-jdbc-standalone.jar.
Make sure to download the standalone version of the Apache Hive JDBC driver to avoid Java library dependency errors, e.g., hive-jdbc-4.0.1-standalone.jar (commonly used driver to connect to Spark).
Done! That was easy, wasn't it? Let's proceed to the next step.

Create ODBC Data Source (DSN) based on ZappySys JDBC Bridge Driver

Step-by-step instructions

To get data from Apache Spark using Microsoft Fabric we first need to create a DSN (Data Source) which will access data from Apache Spark. We will later be able to read data using Microsoft Fabric. Perform these steps:

Download and install ODBC PowerPack.
Open ODBC Data Sources (x64):
Create a User data source (User DSN) based on ZappySys JDBC Bridge Driver:

ZappySys JDBC Bridge Driver
- Create and use User DSN if the client application is run under a User Account. This is an ideal option in design-time, when developing a solution, e.g. in Visual Studio 2019. Use it for both type of applications - 64-bit and 32-bit.
- Create and use System DSN if the client application is launched under a System Account, e.g. as a Windows Service. Usually, this is an ideal option to use in a production environment. Use ODBC Data Source Administrator (32-bit), instead of 64-bit version, if Windows Service is a 32-bit application.
Now, we need to configure the JDBC connection in the new ODBC data source. Simply enter the Connection string, credentials, configure other settings, and then click Test Connection button to test the connection:

ApacheSparkDSN

jdbc:hive2://spark-thrift-server-host:10000

D:\Drivers\JDBC\hive-jdbc-standalone.jar

[]

Use these values when setting parameters:
- Connection string: jdbc:hive2://spark-thrift-server-host:10000
- JDBC driver file(s): D:\Drivers\JDBC\hive-jdbc-standalone.jar
- Connection parameters: []
You should see a message saying that connection test is successful:

Otherwise, if you are getting an error, check out our Community for troubleshooting tips.
We are at the point where we can preview a SQL query. For more SQL query examples visit JDBC Bridge documentation:
ApacheSparkDSN
```
-- Basic SELECT with a WHERE clause
SELECT
    id,
    name,
    salary
FROM employees
WHERE department = 'Sales';
```
```
-- Basic SELECT with a WHERE clause
SELECT
    id,
    name,
    salary
FROM employees
WHERE department = 'Sales';
```
You can also click on the <Select Table> dropdown and select a table from the list.

The ZappySys JDBC Bridge Driver acts as a transparent intermediary, passing SQL queries directly to the JDBC driver, which then handles the query execution. This means the Bridge Driver simply relays the SQL query without altering it.

Some JDBC drivers don't support INSERT/UPDATE/DELETE statements, so you may get an error saying "action is not supported" or a similar one. Please, be aware, this is not the limitation of ZappySys JDBC Bridge Driver, but is a limitation of the specific JDBC driver you are using.
Click OK to finish creating the data source.

Video Tutorial

Installing Microsoft On-premises data gateway

To access and read Apache Spark data in Microsoft Fabric, you will have to download and install Microsoft On-premises data gateway.

There are two types of On-premises data gateways:

Standard Mode
Personal Mode

Standard Mode supports Power BI and other Microsoft Data Fabric services. It fits perfectly for Enterprise solutions as it installs as a Windows Service and also supports Direct Query feature.

Personal Mode, on the other hand, can be configured faster, but is designed more for home users (you cannot install it as a Windows Service and it does not support DirectQuery). You will find a detailed comparison in the link above.

We recommend to go with Personal Mode for a quick POC solution, but use Standard Mode in production environment. You can download On-premises data gateway straight from Microsoft Data Fabric or Power BI web app:

Download Power BI On-premises data gateway

Standard mode

Follow these steps how to configure On-premises data gateway in Standard mode:

Install On-premises data gateway (standard mode) and sign-in:

Use the same email address you use when logging in into your account.
Register a new gateway (or migrate an existing one):
If you are creating a new gateway, name your gateway, enter a Recovery key, and click Configure button:
Once Microsoft gateway is installed, check if it registered correctly:
- Go back to Microsoft Fabric portal
- Click Gear icon on top-right
- And then hit Manage connections and gateways menu item
Continue by clicking On-premises data gateway tab and select Standard mode gateways option from the dropdown:
If your gateway is not listed, the registration may have failed. To resolve this:
- Wait a couple of minutes and refresh Microsoft Fabric portal page
- Restart the machine where On-premises data gateway is installed
- Check firewall settings
Done!

Now we are ready to read data in Microsoft Fabric.

Personal mode

Follow this single step to configure On-premises data gateway in Personal mode:

Install On-premises data gateway (personal mode) and sign-in:

Use the same email address you use when logging in into your account.
Once Microsoft gateway is installed, check if it registered correctly:
- Go back to Microsoft Fabric portal
- Click Gear icon on top-right
- And then hit Manage connections and gateways menu item
Continue by clicking On-premises data gateway tab and select Personal mode option from the dropdown:
If your gateway is not listed, the registration may have failed. To resolve this:
- Wait a couple of minutes and refresh Microsoft Fabric portal page
- Restart the machine where On-premises data gateway is installed
- Check firewall settings
Done!

Now we are ready to read data in Microsoft Fabric.

Loading Apache Spark data into Microsoft Fabric

We created ODBC data source and installed Microsoft On-premises data gateway, and we are ready to load the data into Microsoft Fabric.

Let's roll!

Log in to the Microsoft Fabric Portal.
Select an existing Workspace or create a new one by clicking New workspace (make sure you are in Home section):
Inside your workspace, click the New item button in the toolbar to start creating your data pipeline:
In the item selection window, choose Copy job to open the data ingestion wizard:
In the Choose data source screen, search for odbc and select the Odbc source:
Then enter your ODBC connection string (DSN=ApacheSparkDSN) and select MyGateway from the Data gateway dropdown we configured in the previous step:

DSN=ApacheSparkDSN

DSN=ApacheSparkDSN
Select the table(s) and preview the data you wish to copy from Apache Spark. Once done, click Next:

DSN=ApacheSparkDSN
Choose your Data Destination. You can create a New Fabric item (like a Lakehouse or Warehouse) or select an existing one:

As an example, we will be using Lakehouse as destination type in this article
Choose Full copy to load all data or Incremental copy if you want to load only changed data in subsequent runs:
Review the Column and Table mappings section:
On the summary screen, review your settings. You can optionally enable Run on schedule. Click Save + Run to execute the job:

DSN=ApacheSparkDSN

DSN=ApacheSparkDSN
The job will enter the queue. Monitor the Status column to see the progress:

DSN=ApacheSparkDSN
Wait for the status to change to Succeeded. Your Apache Spark data is now successfully integrated into Microsoft Fabric!
Let's go to our Lakehouse MyLakehouse and double-check if it's true:
It's true, done!

Centralized data access via Data Gateway

In some situations, you may need to provide Apache Spark data access to multiple users or services. Configuring the data source on a Data Gateway creates a single, centralized connection point for this purpose.

This configuration provides two primary advantages:

Centralized data access
The data source is configured once on the gateway, eliminating the need to set it up individually on each user's machine or application. This significantly simplifies the management process.
Centralized access control
Since all connections route through the gateway, access can be governed or revoked from a single location for all users.

	Data Gateway	Local ODBC data source
Simple configuration
Installation	Single machine	Per machine
Connectivity	Local and remote	Local only
Connections limit	Limited by License	Unlimited
Central data access
Central access control
More flexible cost

If you need any of these requirements, you will have to create a data source in Data Gateway to connect to Apache Spark, and to create an ODBC data source to connect to Data Gateway in Microsoft Fabric.

Let's not wait and get going!

Creating Apache Spark data source in Gateway

In this section we will create a data source for Apache Spark in Data Gateway. Let's follow these steps to accomplish that:

Search for gateway in Windows Start Menu and open ZappySys Data Gateway Configuration:
Go to Users tab and follow these steps to add a Data Gateway user:
- Click Add button
- In Login field enter username, e.g., john
- Then enter a Password
- Check Is Administrator checkbox
- Click OK to save
Now we are ready to add a data source:
- Click Add button
- Give Datasource a name (have it handy for later)
- Then select Native - ZappySys JDBC Bridge Driver
- Finally, click OK
ApacheSparkDSN

ZappySys JDBC Bridge Driver
When the ZappySys JDBC Bridge Driver configuration window opens, configure the Data Source the same way you configured it in ODBC Data Sources (64-bit), in the beginning of this article.
Very important step. Now, after creating or modifying the data source make sure you:
- Click the Save button to persist your changes.
- Hit Yes, once asked if you want to restart the Data Gateway service.
This will ensure all changes are properly applied:

Skipping this step may result in the new settings not taking effect and, therefore you will not be able to connect to the data source.

Creating ODBC data source for Data Gateway

In this part we will create ODBC data source to connect to Data Gateway from Microsoft Fabric. To achieve that, let's perform these steps:

Open ODBC Data Sources (x64):
Create a User data source (User DSN) based on ODBC Driver 17 for SQL Server:

ODBC Driver 17 for SQL Server

If you don't see ODBC Driver 17 for SQL Server driver in the list, choose a similar version driver.
Then set a Name of the data source (e.g. Gateway) and the address of the Data Gateway:

GatewayDSN

localhost,5000

Make sure you separate the hostname and port with a comma, e.g. localhost,5000.
Proceed with authentication part:
- Select SQL Server authentication
- In Login ID field enter the user name you used in Data Gateway, e.g., john
- Set Password to the one you configured in Data Gateway
Then set the default database property to ApacheSparkDSN (the one we used in Data Gateway):

ApacheSparkDSN
Continue by checking Trust server certificate option:
Once you do that, test the connection:
If connection is successful, everything is good:
Done!

We are ready to move to the final step. Let's do it!

Accessing data in Microsoft Fabric via Data Gateway

Finally, we are ready to read data from Apache Spark in Microsoft Fabric via Data Gateway. Follow these final steps:

Go back to Microsoft Fabric.
In the Choose data source screen, search for odbc and select the Odbc source:
Then enter your ODBC connection string (DSN=GatewayDSN) and select MyGateway from the Data gateway dropdown we configured in the previous step:

DSN=GatewayDSN

DSN=GatewayDSN
Read the data the same way we discussed at the beginning of this article.
That's it!

Now you can connect to Apache Spark data in Microsoft Fabric via the Data Gateway.

If you are asked for authentication details, use Database authentication or SQL Authentication option and enter credentials you used when configuring Data Gateway, e.g. john and your password.

Troubleshooters & resources (JDBC Bridge Driver)

Below are some useful community articles to help you troubleshoot and configure the ZappySys JDBC Bridge Driver:

How to combine multiple JAR files
Learn how to merge multiple .jar dependencies when your JDBC driver requires more than one file.
How to fix JBR error: “Data lake is not available / Unable to verify trust for server certificate chain”
Resolve SSL or certificate validation issues encountered during JDBC connections.
System Exception: “Java is not installed or not accessible”
Fix Java path or environment issues that prevent the JDBC Bridge from launching Java.
JDBC Bridge Driver disconnect from Java host error
Troubleshoot unexpected disconnection problems between SSIS and the Java process.
Error: Could not open jvm.cfg while using JDBC Bridge Driver
Resolve JVM configuration path errors during driver initialization.
How to enable JDBC Bridge Driver logging
Enable detailed driver logging for better visibility during troubleshooting.
How to pass JDBC connection parameters (not by URL)
Learn how to specify connection properties programmatically instead of embedding them in the JDBC URL.
How to fix JDBC Bridge error: “No connection could be made because the target machine actively refused it”
Troubleshoot firewall or local port binding issues preventing communication with the Java host.
How to use JDBC Bridge options (System Property for Java command line, e.g., classpath, proxy)
Configure custom Java options like classpath and proxy using JDBC Bridge system properties.

Conclusion

In this article we showed you how to connect to Apache Spark in Microsoft Fabric and integrate data without any coding, saving you time and effort. It's worth noting that ZappySys JDBC Bridge Driver allows you to connect not only to Apache Spark, but to any Java application that supports JDBC (just use a different JDBC driver and configure it appropriately).

We encourage you to download Apache Spark Connector for Microsoft Fabric and see how easy it is to use it for yourself or your team.

If you have any questions, feel free to contact ZappySys support team. You can also open a live chat immediately by clicking on the chat icon below.

Download Apache Spark Connector for Microsoft Fabric Documentation