Tableau Calculations¶

Tableau offers multiple calculation types that operate at different levels of granularity. Understanding when to use each type is essential for building accurate and performant dashboards.

Key Facts¶

• Six calculation types: row-level, aggregated, quick table calculations, table calculations, window calculations, LOD
• Row-level calculations run per row; aggregated require aggregation functions
• Table calculations compute over values visible in the current view
• Window calculations are a sub-type of table calculations that operate over a window of visible values
• Quick table calculations are wrappers around window functions applied in 2 clicks
• Parameters are user-controlled variables that modify dashboard behavior

Patterns¶

Calculation Types Overview¶

Date Functions¶

DATEPART('month', [Order Date])    -- extracts integer (3 for March)
DATETRUNC('month', [Order Date])   -- returns first date of period (2023-03-01)
DATEDIFF('day', [Order Date], TODAY())  -- difference between dates
DATENAME('weekday', [Order Date])  -- returns string ("Monday")

Note: Tableau wraps DATETRUNC in DATEBAR by default. Choose "exact date" from field dropdown to see actual date.

String Functions¶

SPLIT([Product Name], '-', 1)      -- split by delimiter, first part
FIND([String], 'word')             -- position of substring
REPLACE([String], 'old', 'new')    -- replace substring
CONTAINS([String], 'word')         -- true/false
STARTSWITH([Category], 'word')     -- true/false (faster than CONTAINS)
REGEXP_MATCH / REGEXP_EXTRACT / REGEXP_REPLACE  -- regex

Performance tip: STARTSWITH is faster than CONTAINS when possible.

Aggregated Calculations¶

SUM([Sales])
AVG([Profit])
MEDIAN([Revenue])
COUNTD([User ID])               -- count distinct
ATTR([Region])                  -- checks if unique; marks * if multiple values
SUM([Revenue] - [Cost]) / SUM([Revenue])  -- margin calculation

Table Calculations¶

Indexing and ranking:

INDEX()                -- row number within partition (1, 2, 3...)
FIRST()                -- offset from first row (0 for first)
LAST()                 -- offset from last row (0 for last)
RANK(expression)       -- rank by value
RANK_UNIQUE(), RANK_DENSE(), RANK_PERCENTILE()

Running totals:

RUNNING_SUM(SUM([Sales]))
RUNNING_AVG(SUM([Profit]))
RUNNING_MAX(SUM([Revenue]))

Lookup and reference:

LOOKUP(SUM([Sales]), -1)       -- value from 1 row earlier
LOOKUP(SUM([Sales]), FIRST())  -- value from first row in partition
PREVIOUS_VALUE(0)              -- previous value or 0 if none

Partition and Address: "Compute Using" settings define which dimension to address (calculate along) and which to partition (reset the calculation).

Window Calculations¶

WINDOW_SUM(SUM([Sales]), -2, 0)    -- rolling 3-period sum
WINDOW_AVG(SUM([Sales]), -2, 0)    -- rolling 3-period average
WINDOW_MAX(SUM([Profit]))           -- max in entire window
WINDOW_MIN(SUM([Profit]))           -- min in entire window
WINDOW_STDEV(SUM([Sales]))         -- std deviation

Offsets: FIRST() = beginning, LAST() = end, 0 = current row. Second and third params are relative start and end.

Quick Table Calculations¶

Applied via right-click measure on shelf -> Quick Table Calculation:

Show last month filter: LAST() = 0 as table calculation filter.

Parameters¶

Create: right-click data panel -> Create Parameter. Types: String, Integer, Float, Boolean, Date.

IF [Sales] > [Parameter: Sales Threshold] THEN 'Above' ELSE 'Below' END

Metric switcher pattern: Parameter + IF/CASE to show different metrics based on dropdown selection.

Logical Functions¶

// IF
IF [Profit] > 0 THEN 'Profitable'
ELSEIF [Profit] = 0 THEN 'Break Even'
ELSE 'Loss'
END

// CASE (cleaner for multiple fixed values)
CASE [Region]
  WHEN 'North' THEN 'Northern Division'
  WHEN 'South' THEN 'Southern Division'
  ELSE 'Other'
END

Gotchas¶

• Aggregation direction matters: configure whether calculation runs across columns or down rows from the "Compute Using" menu
• Quick table calculations are actually window functions - they wrap formulas so you don't write them manually
• Mixing aggregated and non-aggregated fields in a calculation causes errors - all must be at the same level
• String type operations are the slowest; prefer type hierarchy: String > Date > Numeric > Boolean
• DUPLICATE_AS_CROSSTAB reveals the underlying data table beneath a visualization - useful for debugging

See Also¶

• tableau lod expressions - FIXED/INCLUDE/EXCLUDE for cross-granularity
• tableau fundamentals - data connections and filter order
• tableau performance optimization - calculation performance tuning
• powerbi fundamentals - DAX equivalent calculations