Calculated fields are how you create new data in Tableau — string manipulation, date calculations, aggregations, conditionals, and LOD expressions. Here is a complete guide covering the calculation types, common patterns, and the LOD expressions that give Tableau its analytical power.
The quick answer
Calculated fields in Tableau let you create new data from existing fields — string manipulation, date calculations, numeric transformations, conditional logic, and aggregations. They are the primary way you extend the data that is in your source beyond what it natively contains. The most powerful calculated field type — Level of Detail (LOD) expressions — allows you to calculate at a different granularity than the current view, enabling analytics patterns that are difficult or impossible in other tools.
Understanding calculated fields is essential for anyone building Tableau dashboards beyond the most basic level. This guide covers all the calculation types, common patterns, and the LOD expressions that give Tableau its analytical depth.
Calculation types
Tableau has four types of calculations:
**Basic calculations** — row-level calculations that operate on each row of data individually, producing one output per input row. String manipulation, date arithmetic, conditional logic, numeric transformations. These run before aggregation.
**Aggregation calculations** — calculations that aggregate values across multiple rows: SUM, AVG, MIN, MAX, COUNT, COUNTD (count distinct), MEDIAN, PERCENTILE. Aggregations collapse multiple rows to a single value based on the current view's level of detail.
**Level of Detail (LOD) expressions** — the most powerful and distinctive Tableau calculation type. LODs calculate at a specified granularity rather than the view's granularity. FIXED calculates at a specified set of dimensions regardless of the view. INCLUDE adds dimensions to the view's level. EXCLUDE removes dimensions from the view's level.
**Table calculations** — calculations that run on the query result set rather than on source data. RUNNING_SUM, LOOKUP, RANK, WINDOW_AVG, PERCENT_OF_TOTAL, FIRST, LAST, INDEX. Table calculations run after aggregation, on the data already returned to Tableau's query engine.
Basic calculations: patterns and examples
**String functions**: manipulate text data.
LEFT([Field], 3) returns the first 3 characters. UPPER([Product Name]) converts to uppercase. TRIM([Email]) removes leading and trailing spaces. CONTAINS([Description], 'urgent') returns true if the description contains 'urgent'. MID([Order Number], 3, 5) returns 5 characters starting at position 3. REPLACE([City], 'New York', 'NYC') replaces a string value. SPLIT([Full Name], ' ', 1) splits on space and returns the first part (first name).
**Date functions**: calculate with dates.
YEAR([Order Date]), MONTH([Order Date]), DAY([Order Date]) extract date components. DATEDIFF('day', [Order Date], [Ship Date]) calculates the number of days between two dates. DATEADD('month', 3, [Start Date]) adds 3 months to a date. DATETRUNC('month', [Order Date]) truncates a date to the first day of its month — useful for monthly aggregations. TODAY() returns the current date. DATEDIFF('day', [Order Date], TODAY()) calculates how many days ago the order was placed.
**Conditional logic**: branch calculations based on conditions.
IF [Sales] > 100000 THEN 'Large' ELSEIF [Sales] > 50000 THEN 'Medium' ELSE 'Small' END — simple conditional classification.
CASE [Region] WHEN 'East' THEN 'Eastern US' WHEN 'West' THEN 'Western US' ELSE 'Other' END — CASE statement for mapping values.
IIF([Profit] > 0, 'Profitable', 'Unprofitable') — shorthand IF for two-condition cases.
IFNULL([Value], 0) — replace NULL with 0 (or any default). ZN([Value]) is shorthand for IFNULL([Value], 0).
**Numeric calculations**: mathematical transformations.
ROUND([Revenue], 2) rounds to 2 decimal places. ABS([Variance]) returns the absolute value. SQRT([Area]) returns the square root. POWER([Base], [Exponent]) raises to a power. LOG([Value]) returns the natural logarithm. MIN([Budget], [Actual]) returns the smaller of two values (row-level minimum, not view-level).
Aggregation calculations
Aggregation calculations are placed on rows, columns, or marks shelves and define how Tableau aggregates values in the current view.
SUM([Sales]) — sum of all sales in the current partition. AVG([Days to Ship]) — average days to ship. MAX([Order Date]) — most recent order date. MIN([Order Date]) — earliest order date. COUNT([Order ID]) — count of rows. COUNTD([Customer ID]) — count of distinct customer IDs.
**Ratio calculations**: calculated aggregations enable derived metrics.
SUM([Profit]) / SUM([Sales]) — profit margin: sum first, then divide. A common mistake is [Profit] / [Sales] (row-level division, then summed) which produces incorrect results for ratio metrics. For ratios, always aggregate both numerator and denominator first.
Level of Detail (LOD) expressions
LOD expressions are Tableau's most powerful and most frequently misunderstood calculation type. They override the view's level of detail, allowing calculations at a different granularity than what is displayed.
The three LOD types:
**FIXED**: calculates at exactly the specified dimensions, ignoring the view's level of detail.
{FIXED [Customer ID] : SUM([Sales])} — calculates total sales per customer, regardless of how the view is structured. If the view is showing data by Region and Month, this field still contains each customer's total lifetime sales — not sales for that region and month.
Practical use: customer lifetime value (total spend per customer), first order date per customer, segment membership based on all-time behaviour.
**INCLUDE**: adds the specified dimensions to the view's existing level of detail.
{INCLUDE [Order ID] : SUM([Quantity])} — calculates the sum of quantity per order, within whatever dimensions the view already has. If the view is at the Product level, this gives total quantity per order-product combination. Useful for calculations that need a finer granularity than the view to be accurate.
**EXCLUDE**: removes the specified dimensions from the view's level of detail.
{EXCLUDE [Order Date] : SUM([Sales])} — in a view showing sales by Product and Order Date, this calculates total sales by Product only (ignoring date). Useful for calculating "percentage of total by dimension X, ignoring dimension Y."
Common LOD patterns
**Customer first order date**:
{FIXED [Customer ID] : MIN([Order Date])}
Creates a field that, for every row, contains the date of that customer's first order. Use this to segment customers by cohort (join month/year), calculate time since first purchase, or identify new vs. returning customers.
**Sales rank within category**:
RANK(SUM([Sales])) — but this is a table calculation. For rank that persists across different view configurations, combine LOD with RANK: create {FIXED [Product] : SUM([Sales])} first, then rank that field.
**Percentage of total (all-time, not current filter)**:
SUM([Sales]) / {FIXED : SUM([Sales])}
The empty FIXED (no dimensions specified) calculates the grand total across all data, ignoring all dimensions. Dividing the view's SUM by the grand total gives percentage of total. Note: dimension filters still apply to FIXED calculations; context filters do not.
**Average order value vs. current view value**:
SUM([Sales]) / COUNTD([Order ID]) — order value in the current view
versus
{FIXED [Customer ID] : SUM([Sales]) / COUNTD([Order ID])} — average order value per customer, regardless of view structure
Table calculations
Table calculations operate on the result set after Tableau has queried and aggregated the data. They are sequential, directional, and depend on how the view is structured.
RUNNING_SUM(SUM([Sales])) — cumulative sum of sales, running left-to-right (or top-to-bottom) across the view. Useful for running totals and cumulative metrics.
LOOKUP(SUM([Sales]), -1) — returns the value of sales from the previous partition. Useful for period-over-period comparisons: LOOKUP(SUM([Sales]), 0) - LOOKUP(SUM([Sales]), -1) gives the change from the previous period.
RANK(SUM([Sales])) — rank of each partition by sales. RANK_DENSE, RANK_MODIFIED, RANK_PERCENTILE are variants.
WINDOW_AVG(SUM([Sales])) — average of sales across the entire table or window. WINDOW_MAX, WINDOW_MIN, WINDOW_SUM are equivalents for other aggregations.
PERCENT_OF_TOTAL(SUM([Sales])) — each value as a percentage of the total. Equivalent to SUM([Sales]) / WINDOW_SUM(SUM([Sales])) but more concise.
**Table calculation scope**: table calculations can be scoped to apply across specific dimensions. Editing the table calculation (right-click → Edit Table Calculation) specifies "Compute Using" — which dimensions define the partitioning and which define the addressing (direction of calculation). Getting this right is the most common table calculation debugging task.
Performance considerations
**Row-level calculations are fast**: basic calculations that operate row-by-row (string manipulation, date functions, conditional logic) add minimal query overhead.
**LOD expressions add query cost**: FIXED LODs generate a separate subquery in the underlying SQL that Tableau executes. For large tables, multiple FIXED LODs in a single view generate multiple subqueries. Test performance with representative data volumes before publishing.
**Table calculations run in memory**: table calculations run on the data already returned to Tableau, not in the database. For very large result sets, complex table calculations can be slow. Reducing the result set size (using filters, aggregating to a higher level) improves table calculation performance.
**Avoid COUNTD on large tables in live connections**: COUNT DISTINCT is expensive on large tables in live database connections — it requires scanning all values to deduplicate. Where possible, materialise the distinct count as a column in the data model rather than computing it at query time.
For the broader context of Tableau performance — extract tuning, VizQL server configuration, backgrounder optimisation — see Tableau Server performance tuning guide. For dashboard design principles including when to use calculated fields vs. data model transformations, see Tableau dashboard design best practices.
Our Tableau consulting team trains analytics engineers and develops complex calculations for enterprise Tableau deployments. If you are building complex analytics in Tableau and want expert support, book a free 30-minute audit.
A former Microsoft data architect audits your data foundation, identifies your top priorities, and sends you a written plan. Free. No pitch.
Book a Call →