WINDOW

The rows from one position to another in the current row's partition, each position counted from the current row or from an end of the partition.

Group: Window and visual calculations. Compatible with DAX.

Syntax

WINDOW ( <from> [, <from_type>], <to> [, <to_type>] [, <relation or axis>] [, <orderBy>] [, <blanks>] [, <partitionBy>] [, <matchBy>] [, <reset>] )

Parameter	Type	Description
`from`	Whole number	Where the window starts.
`from_type`	Keyword optional	`REL` (when left out): from counts from the current row, negative back; `ABS`: from is a place in the partition, 1 the first, -1 the last.
`to`	Whole number	Where the window ends; that row is in it.
`to_type`	Keyword optional	`REL` (when left out) or `ABS`, as for from.
`relation or axis`	Table optional	The rows it works among: a table of the model or one the query defines. Left out, the values the ORDERBY and PARTITIONBY columns have in the selection (as `ALLSELECTED` gives them). In a visual calculation, an axis such as `ROWS` instead.
`orderBy`	ORDERBY optional	`ORDERBY ( ... )`: how the rows of each partition are ordered. Left out, by all the relation's columns, smallest first.
`blanks`	Keyword optional	Where blank values go in the order: `DEFAULT` (or `KEEP`), the only choice here outside a visual calculation.
`partitionBy`	PARTITIONBY optional	`PARTITIONBY ( ... )`: columns of the relation that split it into groups; the function stays inside the current row's group.
`matchBy`	MATCHBY optional	`MATCHBY ( ... )`: the columns that identify the current row. Left out, all the relation's columns.
`reset`	Reset optional	In a visual calculation only: where the count starts again, as for `PREVIOUS`.

Returns

Table. The relation's rows from the start to the end of the window, both included.

Remarks

WINDOW ( -2, REL, 0, REL, ... ) is the current row and the two before it: a three-row moving window.
WINDOW ( 1, ABS, 0, REL, ... ) runs from the first row to the current one: a running total, year to date within a partition by year.
The window is cut at the partition's edges; when it starts after it ends, or lies wholly outside, it is empty.
With both ends ABS every row of the partition gets the same window.
In a visual calculation the relation can be an axis, as for OFFSET; RANGE and MOVINGAVERAGE are the shortcuts there.

Examples

A moving average and a running total

EVALUATE
SUMMARIZECOLUMNS (
    Sales[Month],
    FILTER ( ALL ( Sales[Month] ), YEAR ( Sales[Month] ) = 2026 && MONTH ( Sales[Month] ) <= 5 ),
    "Liters", [Liters sold],
    "Three-month average",
        AVERAGEX ( WINDOW ( -2, REL, 0, REL, ALLSELECTED ( Sales[Month] ), ORDERBY ( Sales[Month] ) ), [Liters sold] ),
    "So far",
        CALCULATE ( [Liters sold], WINDOW ( 1, ABS, 0, REL, ALLSELECTED ( Sales[Month] ), ORDERBY ( Sales[Month] ) ) )
)
ORDER BY Sales[Month]

Sales[Month]	Liters	Three-month average	So far
2026-01-01	45549	45549	45549
2026-02-01	45985	45767	91534
2026-03-01	45109	45547.6666666667	136643
2026-04-01	48416	46503.3333333333	185059
2026-05-01	52329	48618	237388

Places 2 to the last, by price

EVALUATE
WINDOW (
    2, ABS,
    -1, ABS,
    ALL ( Products[Product], Products[Price per liter] ),
    ORDERBY ( Products[Price per liter], ASC )
)
ORDER BY Products[Price per liter]

Products[Product]	Products[Price per liter]
Pils	38
Cider	52
IPA	64

Neighbours on a table the query defines

Types left out are REL: the shop before, the shop itself and the one after, in order of liters.

DEFINE
    TABLE Shops =
        DATATABLE (
            "Shop", STRING, "Liters", INTEGER,
            { { "A", 50 }, { "B", 80 }, { "C", 80 }, { "D", 20 } }
        )
EVALUATE
ADDCOLUMNS (
    Shops,
    "Neighbours",
        CONCATENATEX ( WINDOW ( -1, 1, Shops, ORDERBY ( Shops[Liters] ) ), Shops[Shop], " " )
)
ORDER BY Shops[Liters], Shops[Shop]

Shops[Shop]	Shops[Liters]	Neighbours
D	20	D A
A	50	D A B
B	80	A B C
C	80	B C

OFFSET, INDEX, RANGE, MOVINGAVERAGE, AVERAGEX.

Last updated 3 November 2026 · OpenBI Studio 0.3.74