How Does Tableau Extract Refresh Work?

When you're building reports in Tableau, you're tapping into one of two main ways of connecting to your data: a live connection or an extract. While a live connection gives you real-time data, it can be slow and strain your database. A Tableau extract - a saved snapshot of your data - offers a high-performance alternative, but it needs to be updated. This article breaks down how a Tableau extract refresh works, the difference between full and incremental refreshes, and how to schedule them so your dashboards are always current.

Graphed

Still Building Reports Manually?

Watch how growth teams are getting answers in seconds — not days.

Watch the Demo

What Exactly is a Tableau Extract?

First, it's helpful to understand what a Tableau extract is and why you'd use one. An extract is a highly compressed, column-oriented snapshot of your data that is stored locally in Tableau's high-performance data engine as a .hyper file.

Unlike a live connection, which queries your database every time you interact with a dashboard (like filtering or sorting), an extract pulls the data into Tableau's memory. This process creates a static subset of your original data source. This leads to three main benefits:

• Better Performance: Querying a local .hyper file is lightning-fast compared to sending queries to a remote database. This means filters apply quicker, visuals load faster, and the entire user experience is smoother, especially with large datasets.
• Offline Access: Since the data is saved with the workbook or published on Tableau Server/Cloud, you can access and interact with your dashboards even when you aren't connected to the original data source.
• Reduced Database Load: By using an extract, you query your production database once during the refresh, not every time a user loads a dashboard. This prevents a high volume of reporting queries from slowing down your core business operations.

The only downside is that the data is only as fresh as the last refresh. That’s why understanding the refresh process is so important.

Graphed

Still Building Reports Manually?

Watch how growth teams are getting answers in seconds — not days.

Watch the Demo

The Refresh Process: Full vs. Incremental

When you "refresh" a Tableau extract, you're telling Tableau to connect to the original data source and update the .hyper file with the new or changed data. There are two primary ways Tableau can do this: a full refresh or an incremental refresh.

Full Refresh

A full refresh does exactly what it sounds like. It erases the entire existing extract and replaces it with a completely new data set from the source. It’s the simplest way to update your data.

• How it works: Tableau connects to your database, runs the query to pull all the data required for the extract, deletes the old .hyper file, and builds a brand-new one from scratch.
• When to use it:

The main drawback is that for very large datasets, a full refresh can take a lot of time and consume a significant amount of resources from both your server or local machine.

Incremental Refresh

An incremental refresh is a smarter, more efficient way to update your data because it only adds new rows to the extract instead of replacing the whole thing. This requires you to tell Tableau how to identify what’s new.

• How it works: When you set up an incremental refresh, you choose a specific column, usually a date (e.g., Order Date), timestamp (e.g., Last Modified), or sequential ID (e.g., row_id). During the refresh, Tableau looks for the highest value currently in that column within the extract. It then queries the source database for only the rows that have a greater value in that column and appends an exact copy of only that row from the source data to your Tableau data source without any interruption to the existing data.
• How to set it up in Tableau Desktop:
• When to use an incremental refresh:

The key thing to be careful about when selecting incremental refresh type is that it never modifies data already in memory - any update to historical records will not show up in your dashboard until a full refresh is forced on your data.

Graphed

Still Building Reports Manually?

Watch how growth teams are getting answers in seconds — not days.

Watch the Demo

Automating Refreshes with a Schedule

Setting up schedules takes the manual work out of keeping your data fresh. You can't schedule refreshes directly within Tableau Desktop, automation is handled by Tableau Server or Tableau Cloud after you've published your workbook or data source.

This automated process is the heart of maintaining reliable, "hands-off" reporting for your organization.

Step-by-Step Guide to Scheduling on Tableau Server/Cloud:

1. Publish Your Source: Start by publishing your dashboard (with the extracts embedded) and/or your separate data source to Tableau Server or Cloud.
2. Set Refresh Options During Publish: When publishing, a dialog box will appear. Here, you can click on the "Schedule" button to choose a refresh as your data source is being published to the server. This helps save a step and make the process easier to set up. Alternatively, you could publish your data source without a schedule setup and do it after the source has been successfully uploaded.
3. Authentication Is Key: In this dialog box, you will be prompted to set your credentials for the data source itself. You will have one of two choices: either embed your password for secure direct access to your database, which is saved on Tableau Server and used for each refresh cycle, or prompt for authentication each time the data source wants to be refreshed by prompting the user for authentication.
4. Create or Assign a Schedule: After publishing, navigate to the data source or workbook on Tableau Server/Cloud.
5. Click the Extract Refreshes tab.
6. Click New Extract Refresh or use one of the existing defined schedules on the server. Site admins often create preset schedules like "Nightly Refresh" or "Hourly Refresh."
7. Configure the schedule according to your needs:
8. Finally, click "Create a Schedule." Your extracts, which have been successfully uploaded from your local machine to your server/Cloud, will now refresh according to the schedule you have just set.

Running Refresher Manually

At times you might not want to wait until your next scheduled refresh to run. You may need fresh data for a meeting, for example, or perform an ad-hoc update. In these situations, a manual refresh can be triggered from both Tableau Desktop and Tableau Server or Cloud.

In Tableau Desktop

Triggering a manual refresh while working locally in Tableau Desktop is simple:

• Right-click on the data source you want to update from your Data Source pane.
• Select the Extract option from the context menu.
• Click on the Refresh option from the dropdown menu to initiate the refresh of the extract.

In Tableau Server or Cloud

Manual refreshes can also be triggered on-demand directly on Tableau Server/Cloud for published workbooks and data sources:

• Navigate to your data source or workbook on the Tableau Server portal.
• Click the three dots (...) to access additional options and choose Refresh Extracts.
• Confirm that you want to run the refresh. Tableau will queue the refresh to run immediately in the server's background process.

Graphed

Still Building Reports Manually?

Watch how growth teams are getting answers in seconds — not days.

Watch the Demo

Common Issues & Troubleshooting Tips

Even with a solid setup, your refreshes can fail sometimes. Here are some common problems and their solutions:

Refresh Failures

• Credential Issues: Expired database passwords are a common cause of refresh failure. Always keep your embedded credentials up-to-date.
• Network Connection Issues: If the Tableau Server can't reach the underlying data source due to network outages, refreshes will fail. Check with your IT team to ensure the connection is stable.
• Underlying Data Structure Changes: Tableau expects set columns to be present when a refresh is run. If columns are renamed or deleted in your source database, the refresh will fail. You'll need to update the data source in Tableau Desktop and republish.

Best Practices for Smooth Refreshes

• Use Incremental Refreshes Whenever Possible: Especially for large, append-only data sets. It's dramatically faster than a full refresh.
• Filter Your Data: Use data source filters to limit the amount of data being brought into the extract. This can speed up the refresh process. For example, if you only need the last two years of data, then set up a filter to exclude older data.
• Hide Unused Fields: Before creating the extract, hide any unused fields. This reduces the extract size and improves the refresh performance.
• Schedule Off-Peak Hours: Run your refreshes during hours when your database is under less load to prevent interference with core operations.
• Set Up Failure Notifications on Tableau Server/Cloud: This will alert you via email if the refresh fails, allowing you to troubleshoot it promptly.

Final Thoughts

Successfully managing Tableau extract refreshes is an essential skill for any data analyst. Understanding the difference between full and incremental refreshes and knowing how to schedule them in Tableau Server or Cloud ensures your dashboards remain both performant and reliable. It automates the task of keeping your reports refreshed, freeing you up to focus on finding insights.

While Tableau is excellent for visualization and managing all data connections, schedules, and refresh logistics can add complexity. A tool like Graphed can help simplify this process by automatically syncing with your data sources in real time, taking care of all the updates in the background. Instead of manually configuring updates, you set questions in plain English, and it builds dashboards that are always up-to-date - no complex setup required.