Translate Tableau calculations to DuckDB SQL.

What is the direct answer to this workbook question?

To migrate a Tableau workbook to DuckDB SQL, map row-level calculated fields to SELECT expressions, convert FIXED, INCLUDE, and EXCLUDE Level of Detail expressions into window functions or CTE subqueries, and translate table calculations into OVER() clauses. You can drop a .twbx file into the xmetric inspector to list its formulas before rewriting.

Why migrate workbook calculations to DuckDB SQL?

Many analysts store business logic inside .twbx workbooks tied to a desktop GUI—Tableau Desktop offers a free local edition for one person (without publishing or sharing) and paid Creator licenses at $75 per user/month, billed annually (as of October 2026). Rewriting those calculations into DuckDB SQL (a free, MIT-licensed in-process analytical SQL database) lets coding agents, Python scripts, CI pipelines, and browser BI tools run the exact same metrics over Parquet and CSV files directly in code.

Drop your .twbx or .twb into the xmetric workbook inspector first to list every data source, worksheet, and calculated field (and spot fields nothing uses or formulas that refer to a missing field) in your browser so you know which formulas are actually used before translating them.

How do common Tableau calculations translate to SQL?

Use scalar SQL for row expressions, explicit grouping or subqueries for LOD grain, and ordered SQL windows for supported table calculations.

How can I translate LOD grain into DuckDB SQL?

Level of Detail expressions are the hardest part of a Tableau-to-SQL migration because an LOD computes an aggregation at a grain different from the worksheet's GROUP BY shelves.

How can a FIXED LOD map to a CTE or window function?

When a worksheet groups by Region and Order Month, but a calculation needs each customer's first purchase date across all history:

// Tableau FIXED LOD
{ FIXED [Customer ID] : MIN([Order Date]) }

-- DuckDB SQL using a window function before the final aggregation
WITH customer_cohorts AS (
  SELECT
    customer_id,
    region,
    order_date,
    sales,
    MIN(order_date) OVER (PARTITION BY customer_id) AS first_order_date
  FROM orders
)
SELECT
  DATE_TRUNC('month', first_order_date) AS cohort_month,
  region,
  SUM(sales) AS total_sales
FROM customer_cohorts
GROUP BY 1, 2;

How can an EXCLUDE LOD map to a windowed denominator?

An EXCLUDE expression removes a dimension that is currently on the view shelves, usually to compute a percent of parent total:

// Tableau EXCLUDE LOD (on a view grouped by Region and Category)
SUM([Sales]) / SUM({ EXCLUDE [Category] : SUM([Sales]) })

-- DuckDB SQL equivalent
SELECT
  region,
  category,
  SUM(sales) AS category_sales,
  SUM(sales) / SUM(SUM(sales)) OVER (PARTITION BY region) AS share_of_region
FROM orders
GROUP BY region, category;

How can an INCLUDE LOD map to a pre-aggregated subquery?

An INCLUDE expression adds a finer dimension (such as Order ID) before applying an outer aggregation (such as AVG across orders in a region):

// Tableau INCLUDE LOD: Average order value per Region
AVG({ INCLUDE [Order ID] : SUM([Sales]) })

-- DuckDB SQL: Group by the included grain first, then average at the view grain
WITH order_totals AS (
  SELECT
    region,
    order_id,
    SUM(sales) AS order_revenue
  FROM orders
  GROUP BY region, order_id
)
SELECT
  region,
  AVG(order_revenue) AS avg_order_value
FROM order_totals
GROUP BY region;

What data does this worked example use?

The initial view uses authored synthetic data, not customer data or a captured agent session.

The example record lists its file, calculation or SQL and expected values. The live browser result is computed by xmetric’s existing engine or workbook inspector. Samples are excluded from own-data activation reporting.