How to write an IF / ELSE calculated field in QuickSight

QuickSight has no IF ... THEN ... ELSE keywords. Conditions go in one function, ifelse, as a list of condition and result pairs followed by a final result for everything else. The first condition that is true wins.

Copy and paste

One condition:

ifelse({Sales} >= 500, 'Large', 'Small')

Several conditions, checked in order:

ifelse(
  isNull({Ship Date}), 'Not shipped',
  dateDiff({Order Date}, {Ship Date}) <= 2, 'On time',
  'Late'
)

AND, OR and NOT, with parentheses:

ifelse(({Region} = 'North' OR {Region} = 'South') AND {Sales} >= 500, 'Focus', 'Other')

Equality against a list of values, with switch:

switch({Status}, 'Shipped', 'Closed', 'Returned', 'Closed', 'Open', 'Active', 'Unknown')

Membership in a list, with in:

ifelse(in({Region}, ['North', 'South']), 'Domestic', 'Export')

On aggregated values (in an analysis):

ifelse(sum({Sales}) >= 1000, 'Above target', 'Below target')

Example

Seven orders, used across the guides of this section:

The shipping status formula returns:

Order 3 has no ship date, so dateDiff would return null and the second condition would not be true. Testing isNull first gives it a label of its own instead of letting it fall into Late.

The rules

• An else is required. ifelse needs exactly one final result. Use NULL if nothing should be returned: the AWS parameter example ends with NULL.
• One type for all results. Every result must be text, or every result a number, or every result a date. Mixing 'None' and 0 is an error; convert with toString or pick one type.
• Parentheses for AND and OR. AND binds tighter than OR. The AWS reference shows how month = 5 OR month < 3 AND year = 2000 and (month = 5 OR month < 3) AND year = 2000 select different rows.
• Text comparisons are case-sensitive. 'north' does not equal 'North'; compare toLower({Region}) = 'north' when the source is inconsistent.
• No BETWEEN. Write {Sales} >= 100 AND {Sales} < 500.
• Comments go between /* and */, and line breaks are allowed, so long conditions can be laid out as above.

Row level or aggregated

A formula with only row-level fields returns a value for every row and can be used as a dimension, like the shipping status. A formula with aggregations is evaluated per group of the visual: in a table by Region, sum({Sales}) >= 1000 is tested per region. One formula cannot mix the two: ifelse(sum({Sales}) >= 1000, {Region}, 'Other') is rejected, and so is any condition that compares sum({Sales}) with a plain field.

An aggregated ifelse cannot be used as a dimension either. To group customers by their total sales, compute the total per customer first; see how to calculate a total per customer regardless of the visual.

ifelse, switch or in

switch is shorter when every condition is “field equals value”, and in when several values share a result. Both are case-sensitive like =. Use ifelse for ranges, nulls and conditions on more than one field.

Checked against the Amazon Quick Sight function reference (October 2026). The example results were recomputed in DuckDB.