2.1 Data Connections & File Formats

Key Takeaways

  • Live connections issue dynamic queries directly to the underlying database in real time, whereas extracts create a static, columnar .hyper snapshot optimized for fast performance and offline analysis.
  • Tableau Workbook files (.twb) contain only visual XML formatting and connection parameters without raw data, while Packaged Workbooks (.twbx) bundle the XML workbook with embedded local data files or extracts into a single compressed archive.
  • Tableau Data Source files (.tds) store connection information and field customizations (aliases, default aggregations, color palettes) to enforce standardized metadata across multiple workbooks without duplicating data.
  • Incremental extract refreshes append only newly added rows based on a designated key column (such as transaction date or sequential ID), avoiding the time and resource overhead of full extract rebuilds.
Last updated: July 2026

2.1 Data Connections & File Formats

Connecting Tableau Desktop to data is the first step in any analytical workflow. Tableau provides flexible connectivity options ranging from local flat files and relational databases to enterprise cloud data warehouses. Choosing the correct connection type—Live versus Extract—and selecting the appropriate Tableau file format (.twb, .twbx, .tds, .tdsx) impacts dashboard query performance, security governance, and file shareability.


Live Connections vs. Extracts

When connecting to a data source, Tableau offers two primary connection modes: Live and Extract.

Live Connections

A Live connection maintains a direct, real-time link to the underlying database or file. Every time an analyst interacts with a worksheet—such as applying a filter, dragging a dimension onto rows, or changing a date range—Tableau generates native database queries (SQL, MDX, or VizQL) and submits them directly to the underlying source system.

  • Primary Advantages:

    • Real-time Data Freshness: Visualizations immediately reflect updates in the underlying database without requiring manual or scheduled refreshes.
    • Leverages Database Processing Power: Complex query processing, indexing, and aggregations are offloaded to high-performance database engines (such as Snowflake, Google BigQuery, or Amazon Redshift).
    • Centralized Data Security: Database-level security rules, user permissions, and row-level access policies remain fully enforced by the source system.
  • Key Drawbacks:

    • Network Latency & Server Load: Frequent user interactions can strain transactional databases (OLTP) and slow down response times if network bandwidth is limited.
    • Dependency on Database Uptime: If the source database goes offline or becomes unresponsive, the Tableau workbook cannot render visual marks.

Extract Connections

An Extract connection takes a static snapshot of data from the source system and compiles it into Tableau’s proprietary, high-performance columnar database format: the Hyper extract engine (.hyper).

  • Primary Advantages:

    • Blazing Fast Performance: Hyper extracts are heavily optimized for analytical queries (OLAP), dramatically accelerating calculation and rendering speeds for slow legacy databases.
    • Offline Accessibility: Analysts can work on dashboards offline without remaining connected to corporate networks or cloud databases.
    • Reduced Database Strain: Queries run against the local or server-hosted .hyper file rather than hammering production databases.
    • Advanced Data Optimization: Extracts allow pre-aggregating data for visible dimensions, filtering out unnecessary historical records, and hiding unused fields prior to extract creation.
  • Key Drawbacks:

    • Data Latency: Visualizations show a snapshot frozen at the exact timestamp of extract creation until a refresh operation is executed.
    • Storage Overhead: Extracts consume disk space on local machines or Tableau Server/Cloud.

Live vs. Extract Comparison Matrix

FeatureLive ConnectionExtract Connection (.hyper)
Query ExecutionNative database queries sent on every interactionHigh-performance in-memory Hyper engine queries
Data FreshnessImmediate / Real-timeLatent (updated via manual or scheduled refresh)
Offline CapabilityNo (requires active database connectivity)Yes (data resides in static extract snapshot)
Database LoadHigh (continuous query traffic)Low (queries target local/server extract)
Best Used ForReal-time operational dashboards, massive cloud warehousesSlow databases, offline work, heavy aggregation viz

Tableau File Types & Formats

Tableau utilizes distinct file extensions depending on whether you are saving workbook layouts, sharing raw data, or standardizing connection metadata across teams.

1. Tableau Workbook (.twb)

A .twb file is an XML-formatted text file that holds all structural configuration instructions for your workbook. It contains worksheet definitions, dashboard layouts, custom formatting rules, calculated fields, parameters, sets, and exact connection pathway details (such as server hostnames or local file directory paths).

CRITICAL EXAM FACT: A .twb file contains NO underlying data. If you email a .twb file linked to an Excel file on your local drive to a colleague, they will receive a "File Not Found" error upon opening it because the raw data file was not included.

2. Tableau Packaged Workbook (.twbx)

A .twbx file is a single compressed Zip archive that bundles the .twb workbook XML file alongside all embedded local data sources (such as .xlsx, .csv, or .hyper files) and custom external assets (like custom shapes or background images).

  • Use Case: Best for sharing completed dashboards with external clients or colleagues who lack direct network access to your internal databases or local drives.

3. Tableau Data Source (.tds)

A .tds file is a lightweight XML file that stores connection metadata and data source customizations. It retains server hostnames, table joins, database credentials (optional), default field properties (such as custom number formatting, default aggregations, and color palettes), field aliases, hierarchies, and calculated fields.

EXAM NOTE: A .tds file does NOT contain raw data rows. It acts as a standardized shortcut configuration so multiple analysts can connect to a database with pre-configured calculated fields and clean field names.

4. Tableau Packaged Data Source (.tdsx)

A .tdsx file is a compressed Zip file that combines a .tds metadata file with any embedded local data files or extract (.hyper) files. It allows data stewards to publish or distribute standardized data packages containing both metadata rules and extract data.

5. Hyper Extract File (.hyper)

The .hyper file is Tableau’s proprietary columnar database extract format. It contains the actual raw data rows converted into an optimized structure for ultra-fast aggregations.

File Format Summary Table

ExtensionFile NameContains Raw Data?Primary Purpose
.twbTableau WorkbookNo (XML structure only)Saving work-in-progress dashboards connected to live databases
.twbxPackaged WorkbookYes (Bundles XML + Local Data/Extracts)Sharing complete workbooks with users outside your local network
.tdsTableau Data SourceNo (XML connection metadata)Standardizing default field properties, aliases, and metrics across teams
.tdsxPackaged Data SourceYes (Bundles .tds + Local Data/Extracts)Distributing standardized data models with embedded offline data
.hyperExtract FileYes (Pure extract data snapshot)Powering high-performance columnar in-memory queries

Data Refresh Strategies: Full vs. Incremental

To maintain data currency in extracts, Tableau supports two distinct refresh operations:

  1. Full Refresh: Rebuilds the .hyper extract from scratch by re-querying the entire source dataset. While thorough, full refreshes can be time-consuming and resource-intensive for multi-million row databases.
  2. Incremental Refresh: Appends only new rows to the existing extract based on a designated sequential key column (such as Transaction_ID or Created_Timestamp).

To configure an incremental refresh, select Extract Data in the Data Source tab, check Incremental refresh, and specify the single column used to identify new rows.

Loading diagram...
Data Connections and File Package Architecture
Test Your Knowledge

Which Tableau file format stores worksheet definitions, dashboard layouts, and connection pathways, but contains NO raw data?

A
B
C
D
Test Your Knowledge

An analyst needs to build a dashboard using a legacy database that executes queries very slowly. However, real-time data is not required. Which connection option should the analyst choose?

A
B
C
D
Test Your Knowledge

What is the primary difference between a Full Extract Refresh and an Incremental Extract Refresh in Tableau?

A
B
C
D