How to Select a Source File from the Publish Tableau Dashboard
When you publish a workbook to Tableau Server or Tableau Cloud, one of the less-discussed but genuinely important steps is choosing which source file backs that published dashboard. Get it right and your colleagues see clean, connected data. Get it wrong and you may publish a workbook pointing at a local extract no one else can refresh — or worse, the wrong data entirely.
This guide breaks down what "source file" means in the Tableau publish workflow, how the selection process works, and which factors determine the right choice for any given situation.
What "Source File" Means in the Tableau Publish Context
In Tableau, a source file refers to the underlying data connection your workbook uses — this could be:
- A live connection to a database, cloud warehouse, or API
- A published data source already hosted on Tableau Server or Cloud
- A local file extract (
.hyperor.tdeformat) embedded in or linked to the workbook - A flat file like an Excel spreadsheet or CSV that the workbook reads directly
When you publish a dashboard, Tableau asks what should happen to that connection. The "select source file" step is where you tell Tableau which version of the data should travel with — or stay connected to — the published workbook.
How the Publish Dialog Works 🗂️
When you go to Server > Publish Workbook in Tableau Desktop, the publish dialog opens. Inside this dialog you'll find a Data Sources section. This is where source file selection happens.
For each data source in your workbook, Tableau gives you options:
Option 1: Embed in Workbook
The data connection credentials and, if applicable, the extract file are bundled directly into the published workbook. Viewers on Tableau Server or Cloud can open and interact with the dashboard without needing their own credentials to the original source.
Option 2: Publish as a Separate Data Source
Tableau publishes the data source independently on the server. Other workbooks can then connect to it. This is common in organizations that want a single source of truth — one centrally governed dataset that multiple dashboards consume.
Option 3: Reference a Previously Published Data Source
If the data source already exists on the server, you can point your workbook at that existing published source rather than uploading a new one. This is where the "select source file" step becomes particularly visible — Tableau may prompt you to browse and confirm which file or connection your workbook should reference.
Selecting a Local File as the Source
If your workbook connects to a local file (Excel, CSV, Access, or a .hyper extract), the publish dialog will often prompt you to locate that file explicitly. Here's what that flow typically looks like:
- Open the publish dialog (Server > Publish Workbook)
- Under Data Sources, click Edit next to the relevant source
- If Tableau flags that the file path is local and unresolvable by the server, it prompts you to browse to the correct file
- Use the file picker to navigate to and select your source file
- Choose whether to embed it or publish it separately
This matters because Tableau Server cannot reach into your local machine's file system after publishing. Whatever file backs the workbook needs to either be embedded at publish time or stored somewhere the server can access (like a shared network path or cloud storage location the server is configured to reach).
Key Variables That Shape the Right Approach
The "correct" way to select your source file isn't universal — it depends on several factors:
| Variable | Why It Matters |
|---|---|
| Data freshness requirements | Embedded extracts are static until manually refreshed; live connections update in real time |
| File location | Local files must be embedded or moved to a server-accessible path |
| Team size and reuse | Shared published data sources reduce duplication across many workbooks |
| Credentials and security | Embedded credentials work for broad access; individual authentication suits controlled environments |
| Extract vs. live | Extracts improve performance on large datasets; live connections reflect changes instantly |
| Server version | Some features (like virtual connections) are only available on newer Tableau Server or Cloud versions |
When Tableau Prompts You to "Find" a Source File
A common point of confusion is when Tableau Desktop shows a warning like "The data source could not be found" or prompts you to browse for the file during publishing. This typically happens when:
- The workbook was built on a different machine and the file path no longer matches
- The original data file was renamed or moved
- The connection was set up using a relative path that doesn't resolve in the publish environment
In these cases, clicking Browse or Edit Connection in the publish dialog lets you navigate to the correct file on your machine. Once selected, Tableau will use that file as the basis for the embedded extract or connection being published. 📁
The Spectrum of Publishing Setups
Different organizations — and individual analysts — use Tableau's publish workflow in meaningfully different ways:
- A solo analyst publishing a one-off dashboard might embed a small Excel file extract and move on
- A BI team maintaining enterprise dashboards typically publishes data sources separately, schedules extract refreshes, and uses certified data sources to ensure consistency
- A developer working with cloud warehouses may never embed anything locally — all connections are live, credentialed at the server level
- A mixed environment might blend embedded extracts for historical snapshots with live connections for current-day data, requiring careful attention to which source file is selected for each layer
Each of these setups uses the same publish dialog, but the decisions made inside it — especially around source file selection — produce very different outcomes in how the dashboard behaves after publishing.
What Source File Selection Doesn't Control
It's worth being clear about scope. Selecting a source file during publish affects data connectivity and packaging — it does not:
- Set permissions on who can view the dashboard (that's managed in project and content permissions)
- Configure extract refresh schedules (set up separately in Tableau Server/Cloud after publishing)
- Determine row-level security rules (those live in the data source itself)
Understanding where source file selection ends and other configuration begins helps avoid troubleshooting rabbit holes later. 🔍
The right choice at the source file step ultimately comes down to how your data is stored, where your server can reach it, how often it needs to refresh, and how many workbooks will rely on it — all factors that vary significantly depending on your specific environment and publishing goals.