Window frame (ROWS BETWEEN, RANGE BETWEEN) in PostgreSQL - Step-by-Step Execution

Concept Flow - Window frame (ROWS BETWEEN, RANGE BETWEEN)

Start Window Function

↓

Partition Data (optional)

↓

Order Data

↓

Define Frame Start and End

↓

Calculate Aggregate over Frame

↓

Return Result for Current Row

↓

Move to Next Row

↓

Repeat Until All Rows Processed

Window functions process rows in partitions and order, then apply an aggregate over a frame defined by ROWS or RANGE boundaries for each row.

Execution Sample

PostgreSQL

SELECT
  id,
  value,
  SUM(value) OVER (ORDER BY id ROWS BETWEEN 1 PRECEDING AND CURRENT ROW) AS sum_2_rows
FROM data;

Calculates a running sum of 'value' for the current row and the one before it, ordered by 'id'.

Execution Table

Step	Current Row (id)	Frame Rows Included (id)	Values in Frame	SUM(value) Result
1	1	[1]	[10]	10
2	2	[1,2]	[10,20]	30
3	3	[2,3]	[20,30]	50
4	4	[3,4]	[30,40]	70
5	5	[4,5]	[40,50]	90
6	6	[5,6]	[50,60]	110
7	7	[6,7]	[60,70]	130
8	8	[7,8]	[70,80]	150
9	9	[8,9]	[80,90]	170
10	10	[9,10]	[90,100]	190
11	11	[10,11]	[100,110]	210
12	12	[11,12]	[110,120]	230
13	13	[12,13]	[120,130]	250
14	14	[13,14]	[130,140]	270
15	15	[14,15]	[140,150]	290
16	16	[15,16]	[150,160]	310
17	17	[16,17]	[160,170]	330
18	18	[17,18]	[170,180]	350
19	19	[18,19]	[180,190]	370
20	20	[19,20]	[190,200]	390
21	21	[20,21]	[200,210]	410
22	22	[21,22]	[210,220]	430
23	23	[22,23]	[220,230]	450
24	24	[23,24]	[230,240]	470
25	25	[24,25]	[240,250]	490
26	26	[25,26]	[250,260]	510
27	27	[26,27]	[260,270]	530
28	28	[27,28]	[270,280]	550
29	29	[28,29]	[280,290]	570
30	30	[29,30]	[290,300]	590
31	31	[30,31]	[300,310]	610
32	32	[31,32]	[310,320]	630
33	33	[32,33]	[320,330]	650
34	34	[33,34]	[330,340]	670
35	35	[34,35]	[340,350]	690
36	36	[35,36]	[350,360]	710
37	37	[36,37]	[360,370]	730
38	38	[37,38]	[370,380]	750
39	39	[38,39]	[380,390]	770
40	40	[39,40]	[390,400]	790
Exit	-	-	-	All rows processed

💡 All rows processed, window function completed

Variable Tracker

Variable	Start	After 1	After 2	After 3	After 4	After 5	Final
Current Row (id)	1	2	3	4	5	6	40
Frame Rows Included (id)	[1]	[1,2]	[2,3]	[3,4]	[4,5]	[5,6]	[39,40]
SUM(value)	10	30	50	70	90	110	790

Key Moments - 3 Insights

Why does the first row only include itself in the frame?

Because the frame is defined as '1 PRECEDING AND CURRENT ROW', and for the first row there is no preceding row, so only the current row is included (see execution_table row 1).

How does the frame move as we go to the next rows?

The frame shifts forward by including the current row and the one before it, dropping the oldest row outside the frame (see execution_table rows 2 to 5).

What happens if there is no ORDER BY in the window function?

Without ORDER BY, the frame cannot be defined in terms of row order, so ROWS BETWEEN or RANGE BETWEEN cannot be applied meaningfully, resulting in an error or default frame (not shown here).

Visual Quiz - 3 Questions

Look at the execution_table at Step 3. Which rows are included in the frame?

A [1, 2, 3]

B [3]

C [2, 3]

D [1, 3]

💡 Hint

Check the 'Frame Rows Included (id)' column at Step 3 in the execution_table.

At which step does the SUM(value) first reach 90?

A Step 4

B Step 5

C Step 6

D Step 3

💡 Hint

Look at the 'SUM(value) Result' column in the execution_table to find when it equals 90.

If the frame was changed to 'ROWS BETWEEN 2 PRECEDING AND CURRENT ROW', how would the frame size change?

A It would include 3 rows per frame

B It would include only 1 row per frame

C It would include all rows in the partition

D It would include no rows

💡 Hint

Consider how '2 PRECEDING AND CURRENT ROW' expands the frame compared to '1 PRECEDING AND CURRENT ROW'.

Concept Snapshot

Window frames define which rows a window function uses for each row.
ROWS BETWEEN counts rows before/after current row.
RANGE BETWEEN uses value ranges in ORDER BY column.
Frames slide as query processes each row.
Use ORDER BY to define row order for frames.
Example: ROWS BETWEEN 1 PRECEDING AND CURRENT ROW sums current and previous row.

Full Transcript

Window frame clauses in PostgreSQL let you control which rows a window function looks at for each row. The process starts by optionally dividing data into partitions, then ordering rows. For each row, a frame is defined using ROWS BETWEEN or RANGE BETWEEN, specifying how many rows or value ranges before and after the current row to include. The window function then calculates its result over this frame. For example, using ROWS BETWEEN 1 PRECEDING AND CURRENT ROW sums the current row's value and the one before it. The frame moves forward as the query processes each row, producing a running calculation. Without ORDER BY, frames cannot be defined properly. This visual trace shows how the frame includes rows and how the sum changes step-by-step.