How to Do a Data Extract in Tableau

A slow and stuttering Tableau workbook is one of the most common frustrations for anyone trying to analyze data. You build a beautiful dashboard, connect it to your database, and then every time you click a filter, you're stuck watching a loading spinner. Your creativity and train of thought come to a grinding halt. This article will walk you through the solution: creating a Tableau Data Extract to make your visualizations fast, responsive, and a pleasure to use.

What is a Tableau Data Extract?

In Tableau, you have two primary ways to connect to your data: Live and Extract.

• A Live connection means that every time you interact with your dashboard - like applying a filter or drilling down into a chart - Tableau sends a query directly to your database. If your database is slow or you're working with millions of rows of data, these queries can take a long time, leading to frustrating delays.
• An Extract connection, on the other hand, takes a snapshot of your data and saves it as a highly compressed, optimized file right on your computer or Tableau Server. This special file type, known as a .hyper file, is built for speed. When you work with an extract, Tableau queries this local file instead of the original database, making your interactions almost instantaneous.

Think of it like painting. A live connection is like painting a landscape while looking out a window. What you see is always current, but if something obstructs your view - a slow network connection or the scene is overly complex - a massive database can slow you down. An extract is like taking a high-resolution photograph of that landscape first. You can then paint from the photo quickly and efficiently, without any interruptions, and work on it from anywhere.

Why Use a Tableau Data Extract? (The Big Benefits)

Switching from a live connection to an extract offers much more than just a speed boost. Here are the main advantages that make it a go-to choice for most Tableau developers.

1. Incredible Performance Improvements

This is the most significant benefit. Because .hyper files are column-based and highly compressed, Tableau can retrieve the data it needs for your visualizations much faster than it could from a traditional row-based database. Simple dashboards can load in seconds instead of minutes, and complex filters that used to cause long waits can now apply in a flash. This fluid interactivity is essential for effective data exploration and analysis.

2. Offline Access and Portability

Since an extract is a self-contained file, you can continue working on your dashboards even when you're not connected to the internet or your company's network. This is a game-changer for anyone who travels or works remotely. You can package your workbook (.twbx) with the extract included and send it to a colleague, who can open and interact with it fully without needing access credentials to the original data source.

3. Reduced Load on Your Database

When multiple users are using a live connection on a heavily used dashboard, they are all hitting the database with queries simultaneously. This can place a significant strain on your data warehouse or transactional database, potentially slowing down critical business operations. Using extracts shifts that analytical workload away from the central database, keeping it available for other important tasks.

4. Additional Functionality

Certain calculations and features in Tableau are only available when you're using a data extract. For instance, functions like COUNTD (Count Distinct) can perform much better with an extract. Features like forecasting and some table calculations are also optimized for, or sometimes exclusively available with, extracts.

Creating Your Data Extract: A Step-by-Step Guide

Making an extract in Tableau is straightforward. Just follow these steps on the Data Source page of your workbook.

Step 1: Connect to Your Data Source

Start by opening Tableau and connecting to your data as you normally would. This could be anything from an Excel file to Google Sheets or a SQL database like BigQuery or Snowflake. Drag the table or tables you need onto the canvas to set up your data model.

Step 2: Switch From Live to Extract

In the top right corner of the Data Source screen, you’ll see two options under "Connection": Live and Extract. By default, Tableau usually selects "Live." Simply click on the radio button next to Extract to make the switch.

That's the core action! The next time you save your work or navigate to a sheet, Tableau will prompt you to save the extract file.

Step 3: Edit Your Extract (The Most Important Step!)

This is where you can make your extract even more efficient. To the right of the "Extract" radio button, you'll see a blue link that says "Edit." Clicking this opens the Extract Data dialog box, giving you several powerful options to optimize your data snapshot:

• Data Storage: Here, you can have Tableau store your data in one table (Logical Table) or multiple physical tables that still retain their joins. The "Logical Table" option often provides the best performance for most use cases.
• Filters: You don't always need to extract all your data. Use the "Add..." button to implement a filter on your extract. For example, if you're only analyzing sales from the last two years, you can add a filter on the date field to exclude older, irrelevant data. This dramatically reduces the size of your extract file and speeds everything up even more.
• Aggregate Data for Visible Dimensions: This is a powerful optimization feature. It rolls up your data to the level of detail displayed in your visualization. For instance, if your data contains every single transaction by the second, but your charts only ever show sales by day, you can check this box to aggregate the data to the daily level. This can reduce millions of rows to just a few thousand, resulting in a tiny, incredibly fast extract. For more granular control, you can roll up using a specific date level, like "Years" or "Months."
• Number of Rows: You can choose to extract all rows or just a sample (e.g., the top 10,000 rows). Sampling is handy when you're in the early stages of building a dashboard with a massive dataset and just need a representative chunk to work with quickly.

Step 4: Save Your Extract (.hyper file)

Once you've set your extract options and you move to a worksheet tab, Tableau will prompt you to save the extract file. Choose a location on your computer, give it a name, and save it. Tableau will then process your data based on your settings and create the .hyper file.

Step 5: How To Refresh Your Extract

An extract is a snapshot in time. To keep your dashboard up-to-date with the latest data, you need to refresh it.

• Full vs. Incremental Refresh
• Manual Refresh (in Tableau Desktop)
• Scheduled Refreshes (on Tableau Server/Cloud)

Best Practices for Working with Extracts

To get the most out of your extracts, keep these simple tips in mind:

1. Be Strategic with Filters: Always add filters to your extract to exclude any data that you know you won't need for your analysis. The smaller the extract, the faster your workbook will be. Every column and row you exclude makes a difference.
2. Hide Unused Fields: After connecting to your data but before creating the extract, you have the option in the Data Source Tab to select and hide specific fields from being included in your visualization workspace. In the data source grid, you can select multiple columns, right-click, and choose "Hide." Tableau won't include these hidden fields in the extract, further reducing file size and making the extract leaner.
3. Use Aggregation Whenever Possible: If your dashboards report on monthly trends, there's no need to store daily or hourly data in your extract. Aggregate your data to the highest level that still supports your analysis.
4. Schedule Refreshes During Off-Hours: If you are using Tableau Server or Cloud, plan your extract refresh schedules for times when system usage is low, such as overnight or on weekends. This minimizes the impact of your operations.

Final Thoughts

Creating a Tableau Data Extract is a fundamental skill for anyone serious about building high-performance dashboards. You move past the limitations of live database connections, creating an optimized and portable data snapshot that unlocks faster analysis, offline capability, and a far better experience for yourself and your audience.

While Tableau extracts are a powerful way to accelerate the performance of a single data source, the initial challenge often lies in an even earlier step: getting all of your disconnected marketing and sales data into one place. This is where we built Graphed to help. We automate the entire process by connecting directly to platforms like Salesforce, HubSpot, Google Analytics, and Facebook Ads, letting you build real-time dashboards simply by describing what you want to see in plain English. No more wrestling with extracts or live connections - just connect your accounts, ask questions, and get instant answers.