How to Create a Hyper File in Tableau

Slow-loading Tableau dashboards are frustrating for you and anyone watching. If you've ever waited for a view to render while clicking around, you've felt the pain of an inefficient data connection. Luckily, Tableau has a powerful, built-in solution designed specifically for this problem: the Hyper file format. This guide will walk you through exactly what .hyper files are, why they are a game-changer for performance, and how you can create and manage them to make your dashboards faster than ever.

Graphed

Still Building Reports Manually?

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

Watch the Demo

What is a Tableau Hyper File, Anyway?

In simple terms, a Tableau Hyper file (with a .hyper extension) is a high-performance, compressed snapshot of your data. Think of it as a supercharged local copy. Instead of Tableau sending a query over a network to your live database every time you filter, sort, or interact with a dashboard, it queries this highly optimized local file.

This technology is more than just a file format, it's Tableau's proprietary in-memory data engine. It's engineered from the ground up to handle massive datasets (billions of rows) with incredible speed, processing queries much faster than most traditional databases. It intelligently compresses data and organizes it in a columnar format, allowing Tableau to read only the data it needs for a specific visualization.

When you connect to your data in Tableau, you have two primary options:

• Live Connection: Each interaction sends a new query directly to the source database. This is great for real-time data but can be slow depending on the database performance, network latency, and query complexity.
• Extract Connection: Tableau takes a snapshot of the data and saves it as a .hyper file. All interactions then query this local, optimized file, leading to significantly faster performance.

The "Why": Key Benefits of Using .hyper Files

Switching from a live connection to a Hyper extract isn't just a technical tweak, it provides tangible benefits that vastly improve both the developer and the end-user experience.

Blazing-Fast Dashboard Performance

This is the most significant benefit. Since the data is stored locally in an optimized format, dashboards load in seconds, not minutes. Interactions like changing filters, drilling down into data, or sorting a large table become almost instantaneous. A report that took 45 seconds to refresh with a live connection could easily load in under 5 seconds using an extract. This responsiveness encourages users to explore the data more deeply instead of giving up out of frustration.

Graphed

Still Building Reports Manually?

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

Watch the Demo

Reduced Load on Your Database

Every time a user interacts with a dashboard using a live connection, it puts a strain on your source database. With dozens or hundreds of users working with dashboards, this can create a significant performance bottleneck for the entire organization (and an unhappy database administrator). Using an extract offloads all of that query work from the production database to the .hyper file, keeping your operational systems running smoothly.

Offline Data Access

Because extracts save a copy of the data directly within your Tableau workbook (when saved as a .twbx), you can work on your analysis without an internet connection. This is perfect for working on a plane, from a coffee shop, or from a home office with an unreliable network. You can build charts, create calculations, and design your entire dashboard offline, then publish it when you reconnect.

Greater Data Portability and Sharing

Packaging a workbook with a Hyper extract makes sharing your analysis incredibly easy. When you save your workbook as a Packaged Workbook (.twbx), the .hyper file is bundled inside. You can then email this single file to a colleague, and they can open it and interact with all the data without needing access credentials for your original database. It's a self-contained analytics package.

Step-by-Step: Creating Your First Hyper File in Tableau

Creating an extract is a straightforward process right within the Tableau Desktop interface. Let's walk through it from start to finish.

Step 1: Connect to Your Data Source

In Tableau Desktop, start by connecting to your data. This can be anything from a simple Excel or flat file to cloud applications like Salesforce or a corporate SQL database. On the 'Connect' pane, select your data source and provide your credentials if necessary.

Step 2: Set Up Your Data Model

Once connected, you'll be on the 'Data Source' tab. Drag the tables you need for your analysis onto the canvas. Tableau will automatically try to create relationships between them based on common field names. You can edit these relationships or create joins just as you normally would.

Pro Tip: Before creating your extract, hide any fields (columns) you know you won't need for your analysis. You can do this by selecting the columns in the data preview pane, right-clicking, and choosing 'Hide'. A leaner data source will result in a smaller and faster extract.

Graphed

Still Building Reports Manually?

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

Watch the Demo

Step 3: Select the 'Extract' Connection Type

This is the most critical step. In the top-right corner of the Data Source page, you'll see two radio buttons: Live and Extract. Simply click on Extract to switch. This tells Tableau that you want to create a .hyper file instead of maintaining a live connection.

Step 4: Configure the Extract (Optional but Highly Recommended)

Beside the 'Extract' option, you'll see a link that says 'Edit...'. Clicking this opens the Extract Data dialog box, where you can fine-tune your snapshot for optimal performance. Here are the key options:

Storage

Under 'Data Storage', you can choose between 'Logical Tables' and 'Physical Tables'. For most situations, the default of 'Logical Tables' is the best choice, as it maintains the normalized structure of your data model and provides more flexibility.

Filters

This is one of the most powerful features for optimizing extracts. Click 'Add...' to open the Filter dialog. You can use this to limit the amount of data being pulled into your snapshot. For example, instead of extracting all 20 years of sales data, you could add a filter on the 'Order Date' field to only include the last 3 years. This can drastically reduce the size of the extract and the time it takes to create and refresh it.

Aggregation

The 'Aggregate data for visible dimensions' option can create even smaller, faster extracts. This summarizes your data to the level of detail specified by your dimensions. For instance, rather than storing every individual transaction, you could roll the data up to 'monthly sales per product category'. The trade-off is that you lose the ability to drill down to the row-level detail, so only use this if your analysis doesn’t require it.

Step 5: Create the Extract File

Once you've configured your settings, simply navigate to any sheet in your workbook (e.g., click on 'Sheet 1'). At this moment, Tableau will prompt you to save the extract file. Choose a location on your computer, give the file a name (it will be saved as a .hyper file), and click 'Save'.

Tableau will then run the necessary queries against your data source, apply your filters and aggregations, and create the optimized file. Voila! You have successfully created a Hyper extract. You'll notice the icon next to your data source in the 'Data' pane has changed from a single cylinder (live) to a double cylinder (extract).

Graphed

Still Building Reports Manually?

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

Watch the Demo

Keeping it Current: Refreshing Your .hyper File

An extract is a snapshot in time, so your next question is likely, "How do I get updated data?" Tableau gives you two great ways to keep your data fresh.

Manual Refresh in Tableau Desktop

While working in Tableau Desktop, you can refresh the extract at any time. Simply right-click on the data source in the 'Data' pane, go to Extract, and select Refresh. This triggers a full refresh, where Tableau deletes the old data and rebuilds the extract from scratch with the latest information from the source.

Scheduled Refreshes on Tableau Server/Cloud

Herein lies the real power of automation. When you publish a workbook that uses an extract to Tableau Server or Tableau Cloud, you can put that extract on a refresh schedule. You can set it to update automatically every hour, once a day at 5 AM, or every Monday morning. Users will always see fresh data in their dashboards without anyone having to manually intervene.

Full Refresh vs. Incremental Refresh

In your extract settings (or on Tableau Server), you can choose between a full or incremental refresh.

• Full Refresh: The default behavior. It erases the existing extract and rebuilds it with all the data.
• Incremental Refresh: This option only adds new rows since the last refresh. It requires a column in your data that uniquely identifies new records, such as an unchanging Order ID or a Created Date timestamp. For datasets with millions or billions of rows, an incremental refresh is dramatically faster as it only processes the new data.

Final Thoughts

Leveraging Tableau's .hyper extracts is a fundamental skill that transforms slow, clunky workbooks into fast, interactive dashboards. By creating an optimized snapshot of your data, you reduce the load on your databases, enable offline access, and deliver a vastly superior analytical experience for your audience.

Moving from live connections to optimized extracts is a fantastic way to speed up your reporting workflow. At Graphed, we are obsessed with taking this a step further by automating the entire analysis process. Instead of manually configuring data sources and building charts, you can connect your platforms like Google Analytics or Salesforce in a few clicks, and then create entire dashboards just by describing what you need in plain English. We designed it so you spend less time wrestling with data prep and more time discovering insights.