1. Introduction to T-sql Window functionS
Kathi Kellenberger
@auntkathi

2. Agenda What are T-SQL window functions? 2005 Ranking functions
Window aggregates
2012 Enhancements
Accumulating window aggregates
Framing
Offset functions
Statistical functions
Agenda

3. What are window functions?
Not OS, based on ANSI-SQL standards
Function performs over a SET (window of the results)
How to identify a window function
The OVER clause defines the window
Where to put window functions
SELECT and ORDER BY clauses only
Important!
The FROM, WHERE, GROUP BY and HAVING clauses operate BEFORE the window functions
Partitions NOT the same as GROUP BY

4. Ranking functions Available since 2005
ROW_NUMBER()
A unique # over the window
RANK()
Deals with ties
DENSE_RANK()
NTILE()
Divide rows into buckets
ORDER BY is required in OVER clause

5. ROW_NUMBER() example CustomerID OrderID Total
ROW_NUMBER() OVER(PARTITION BY CustomerID Order by OrderID) 1
101
100
105
2000 2
106
300 3
102 40
103 11
104
432 4
107
674
109
234
110
889 5
108 76
111

6. Demo 1
Ranking functions

7. Window aggregate functions
Your favorite aggregates with no GROUP BY
Add aggregate function to a non-aggregate query
Calculate over the window: the entire result set or partition
No ORDER BY
Partition by OR use the Empty Over Clause ∑

8. Window Aggregate example
CustomerID
OrderID
TotalDue
SUM(TotalDue) OVER(Partition by CustomerID) 1
101
100
2400
102
2000
103
300 2
104 40 51
105 11 3
106
432 4
107
674
1797
108
234
109
889 5
110 76
310
111

9. Demo 2
Window aggregates

10. 2012 enhancements
Accumulation aggregates Even further define the window with FRAMING Eight new functions Offset Statistical

11. Accumulating aggregate example
CustomerID
OrderID
TotalDue
SUM(TotalDue) OVER(Order by OrderID) 1
101
100 2
102 40
140
103 11
151 3
104
432
583
105
2000
2583
106
300
2883 4
107
674
3557 5
108 76
3633
109
234
3867
110
889
4756
111
4990

12. Accumulating aggregates
Demo 3
Accumulating aggregates

13. FRAMING: ROWS and RANGE
Further define the window Each row has its own window Supported for specific function types

14. Framing: Vocabulary Term Used for ROWS
Positional operator used to define the frame
RANGE
Logical operator used to define the frame. Not fully supported. The default if ROWS is not specified
UNBOUNDED PRECEDING
The first row of the partition
UNBOUNDED FOLLOWING
The last row of the partition
CURRENT ROW
The row where the calculation is being performed

15. FRAMING ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW
ROWS UNBOUNDED PRECEDING

16. FRAMING ROWS BETWEEN CURRENT ROW AND UNBOUNDED FOLLOWING 1 2 3 4 5 6 78 9 10 11 12 13 14 15
ROWS BETWEEN CURRENT ROW AND UNBOUNDED FOLLOWING

17. FRAMING ROWS BETWEEN 2 PRECEDING AND CURRENT ROW 1 2 3 4 5 6 7 8 9 1011 12 13 14 15
ROWS BETWEEN 2 PRECEDING AND CURRENT ROW

18. Demo 4
Framing

19. OFFSET Functions
LAG() Grab a column from a previous row LEAD() Grab a column from a later row FIRST_VALUE() Grab a column from the first row Supports framing LAST_VALUE() Grab a column from the last row Supports framing -- be careful!!

20. LAG example CustomerID OrderID TotalDue
LAG(TotalDue) OVER(ORDER BY OrderID) 1
101
100
NULL 2
102 40
103 11 3
104
432
105
2000
106
300 4
107
674 5
108 76
109
234
110
889
111

21. LEAD example CustomerID OrderID TotalDue
LEAD(TotalDue) OVER(ORDER BY OrderID) 1
101
100 40 2
102 11
103
432 3
104
2000
105
300
106
674 4
107 76 5
108
234
109
889
110
111
NULL

22. First_VALUE example CustomerID OrderID TotalDue
FIRST_VALUE (TotalDue) OVER(ORDER BY OrderID) 1
101
100 2
102 40
103 11 3
104
432
105
2000
106
300 4
107
674 5
108 76
109
234
110
889
111

23. LAST_VALUE example CustomerID OrderID TotalDue
LAST_VALUE(TotalDue) OVER(ORDER BY OrderID) 1
101
100
234 2
102 40
103 11 3
104
432
105
2000
106
300 4
107
674 5
108 76
109
110
889
111

24. Demo 5
Offset functions

25. Statistical functions
PERCENT_RANK() Relative rank. “My score is better than 90% of the scores” CUME_DIST() Cumulative distribution over a group of rows. “My score is at 90%” PERCENTILE_DISC() Computes a specific percentile PERCENTILE_CONT() Computes an exact percentile

26. PERCENT_RANK example

27. CUME_DIST example

28. PERCENTILE_DISC example

29. PERCENTILE_CONT example

30. Statistical functions
Demo 6
Statistical functions

31. Resources My book: Expert T-SQL Window Functions
Itzik Ben-Gan’s book: Microsoft SQL Server High-Performance T-SQL Using Window Functions
My Pluralsight course