Firebird 3 Windows Functions

Firebird 3 Windowing Functions


Author: Philippe Makowski IBPhoenix
Email: pmakowski@ibphoenix
Licence: Public Documentation License
Date: 2011-11-22

Philippe Makowski - IBPhoenix - 2011-11-22


What are Windowing Functions?
• Similar to classical aggregates but does more!
• Provides access to set of rows from the current row
• Introduced SQL:2003 and more detail in SQL:2008
• Supported by PostgreSQL, Oracle, SQL Server, Sybase and DB2
• Used in OLAP mainly but also useful in OLTP
• Analysis and reporting by rankings, cumulative aggregates



Windowed Table Functions
• Windowed table function
• operates on a window of a table
• returns a value for every row in that window
• the value is calculated by taking into consideration values from the set of
rows in that window
• 8 new windowed table functions
• In addition, old aggregate functions can also be used as windowed table
functions
• Allows calculation of moving and cumulative aggregate values.



A Window
• Represents set of rows that is used to compute additionnal attributes
• Based on three main concepts
• partition
• specified by PARTITION BY clause in OVER()
• Allows to subdivide the table, much like GROUP BY clause
• Without a PARTITION BY clause, the whole table is in a single partition
• order
• defines an order with a partition
• may contain multiple order items
• Each item includes a value-expression
• NULLS FIRST/LAST defines ordering semantics for NULL
• this clause is independant of the query's ORDER BY clause
• frame (Firebird don't implement frame yet)



Window function syntax in Firebird
<window function> ::=
<window function type> OVER <window specification>

<window function type> ::=
<rank function type> <left paren> <right paren>
| ROW_NUMBER <left paren> <right paren>
| <aggregate function>
| <lead or lag function>
| <first or last value function>
| <nth value function>

<rank function type> ::=
RANK
| DENSE_RANK



<first or last value function> ::=
<first or last value> <left paren> <value expression> <right paren>

<first or last value> ::=
FIRST_VALUE | LAST_VALUE

<nth value function> ::=
NTH_VALUE <left paren> <value expression> <comma> <nth row>
<right paren>

<nth row> ::=
<simple value specification>
| <dynamic parameter specification>
(must be an integer >= 1)



Window specification syntax in Firebird
<window specification> ::=
<left paren> <window specification details> <right paren>

<window specification details> ::=
[ <window partition clause> ]
[ <window order clause> ]

<window partition clause> ::=
PARTITION BY <window partition column reference or expression list>

<window partition column reference or expression list> ::=
<window partition column reference or expression>
[ { <comma> <window partition column reference or expression> }... ]

<window partition column reference or expression> ::=
<column reference or expression> [ <collate clause> ]

<window order clause> ::=
ORDER BY <sort specification list>



About order
The window order clause is to calculate its values, and has nothing to do with the main
order.

select emp_no, salary,
sum(salary) over (order by salary) cum_salary,
sum(salary) over (order by salary desc) cum_salary_desc
from employee order by emp_no;
EMP_NO SALARY CUM_SALARY CUM_SALARY_DESC
======= ===================== ===================== =====================
2 105900.00 1990493.02 113637875.00
4 97500.00 1680929.02 113939039.00
28 22935.00 22935.00 115522468.02
121 99000000.00 115522468.02 99000000.00
145 32000.00 113210.00 115441258.02



About order
If the main query doesn't have a order by but a window have, the query will be ordered
accordingly to the last window order.

sum(salary) over (order by salary) cum_salary,
sum(salary) over (order by salary desc) cum_salary_desc
from employee;
EMP_NO SALARY CUM_SALARY CUM_SALARY_DESC
======= ===================== ===================== =====================
121 99000000.00 115522468.02 99000000.00
118 7480000.00 16522468.02 106480000.00
110 6000000.00 9042468.02 112480000.00
134 390500.00 3042468.02 112870500.00



About order
sum(salary) over (order by salary desc) cum_salary_desc,
sum(salary) over (order by salary) cum_salary
from employee;
EMP_NO SALARY CUM_SALARY_DESC CUM_SALARY
======= ===================== ===================== =====================
28 22935.00 115522468.02 22935.00
109 27000.00 115499533.02 49935.00
65 31275.00 115472533.02 81210.00
145 32000.00 115441258.02 113210.00



Set Functions as Window Functions
The OVER clause turns a set function into a window function
• Aggregated value is computed per current row window

select emp_no, dept_no, salary,
avg(salary) over (partition by dept_no) as dept_avg
from employee;
EMP_NO DEPT_NO SALARY DEPT_AVG
======= ======= ===================== =====================
12 000 53793.00 133321.50
105 000 212850.00 133321.50
127 100 44000.00 77631.25
85 100 111262.50 77631.25
34 110 61637.81 65221.40
61 110 68805.00 65221.40
110 115 6000000.00 6740000.00
118 115 7480000.00 6740000.00



Set Functions as Window Functions
Who are the highest paid relatively compared with the department average?

avg(salary) over (partition by dept_no) as dept_avg,
salary - avg(salary) over (partition by dept_no) as diff
from employee
order by diff desc;
EMP_NO DEPT_NO SALARY DEPT_AVG DIFF
======= ======= ===================== ===================== =====================
118 115 7480000.00 6740000.00 740000.00
105 000 212850.00 133321.50 79528.50
107 670 111262.50 71268.75 39993.75
2 600 105900.00 66450.00 39450.00
85 100 111262.50 77631.25 33631.25
4 621 97500.00 69184.87 28315.13
46 900 116100.00 92791.31 23308.69
9 622 75060.00 53409.16 21650.84



Built-in Windowing Functions
• RANK () OVER ...
• DENSE_RANK () OVER ...
• LAG () OVER ...
• LEAD () OVER ...
• ROW_NUMBER () OVER ...
• FIRST_VALUE () OVER ...
• LAST_VALUES () OVER ...
• NTH_VALUE () OVER ...



Built-in Windowing Functions : row_number
Returns number of the current row

row_number() over (order by salary desc nulls last) as row_num
from employee;
EMP_NO DEPT_NO SALARY ROW_NUM
======= ======= ===================== =====================
121 125 99000000.00 1
118 115 7480000.00 2
110 115 6000000.00 3
134 123 390500.00 4
105 000 212850.00 5
46 900 116100.00 6
85 100 111262.50 7
107 670 111262.50 8
141 121 110000.00 9

row_number() always incremented values independent of frame



Built-in Windowing Functions : rank
Returns rank of the current row with gap

rank() over (order by salary desc nulls last) as rank
from employee;
EMP_NO DEPT_NO SALARY RANK
======= ======= ===================== =====================
121 125 99000000.00 1
118 115 7480000.00 2
110 115 6000000.00 3
134 123 390500.00 4
105 000 212850.00 5
46 900 116100.00 6
85 100 111262.50 7
107 670 111262.50 7
141 121 110000.00 9

rank() OVER(empty) returns 1 for all rows, since all rows are peers to each other



Built-in Windowing Functions : dense_rank
Returns rank of the current row without gap

dense_rank() over (order by salary desc nulls last)
from employee;
EMP_NO DEPT_NO SALARY DENSE_RANK
======= ======= ===================== =====================
121 125 99000000.00 1
118 115 7480000.00 2
110 115 6000000.00 3
134 123 390500.00 4
105 000 212850.00 5
46 900 116100.00 6
85 100 111262.50 7
107 670 111262.50 7
141 121 110000.00 8

dense_rank() OVER(empty) returns 1 for all rows, since all rows are peers to each
other



Built-in Windowing Functions : rank, dense_rank,
row_number
rank() over (order by salary desc nulls last),
dense_rank() over (order by salary desc nulls last),
row_number() over (order by salary desc nulls last)
from employee;
EMP_NO DEPT_NO SALARY RANK DENSE_RANK ROW_NUMBER
======= ======= ============= ======= ============= =============
121 125 99000000.00 1 1 1
118 115 7480000.00 2 2 2
110 115 6000000.00 3 3 3
134 123 390500.00 4 4 4
105 000 212850.00 5 5 5
46 900 116100.00 6 6 6
85 100 111262.50 7 7 7
107 670 111262.50 7 7 8
141 121 110000.00 9 8 9



Missing Built-in Windowing Functions : percent_rank
Returns relative rank of the current row: (rank - 1) / (total rows - 1)
can be emulated with rank() and count()

cast((rank() over (order by salary asc))-1 as double precision)
/ ( count(*) over () -1) as percent_rank
from employee
EMP_NO DEPT_NO SALARY PERCENT_RANK
======= ======= ===================== =======================
28 120 22935.00 0.000000000000000
109 600 27000.00 0.02439024390243903
...
144 672 35000.00 0.1219512195121951
114 623 35000.00 0.1219512195121951
138 621 36000.00 0.1707317073170732
...
118 115 7480000.00 0.9756097560975610
121 125 99000000.00 1.000000000000000



Missing Built-in Windowing Functions : cum_dist
Returns relative rank; (# of preced. or peers) / (total row)
can be emulated with count()

cast((count(*) over (order by salary asc)) as double precision)
/ ( COUNT(*) over () ) as cum_dist
from employee
EMP_NO DEPT_NO SALARY CUM_DIST
======= ======= ===================== =======================
28 120 22935.00 0.02380952380952381
109 600 27000.00 0.04761904761904762
...
144 672 35000.00 0.1666666666666667
114 623 35000.00 0.1666666666666667
138 621 36000.00 0.1904761904761905
...
118 115 7480000.00 0.9761904761904762
121 125 99000000.00 1.000000000000000



Built-in Windowing Functions :
first_value,nth_value,last_value

Warning
Note that first_value, last_value, and nth_value consider only the rows within
the "window frame", which by default contains the rows from the start of the
partition through the last peer of the current row.
And Firebird today don't allow you to change the frame definition.
This is likely to give unhelpful results for nth_value and particularly last_value.



Built-in Windowing Functions : first_value
Returns value of the first row in a window

first_value(salary) over (order by salary desc nulls last) overall_max,
first_value(salary) over (partition by dept_no order by salary desc nulls last) dept_max,
first_value(salary) over (order by salary asc nulls last) overall_min,
first_value(salary) over (partition by dept_no order by salary asc nulls last) dept_min
from employee;
EMP_NO DEPT_NO SALARY OVERALL_MAX DEPT_MAX OVERALL_MIN DEPT_MIN
======= ======= ============ ============ ============ ============ ============
12 000 53793.00 99000000.00 212850.00 22935.00 53793.00
105 000 212850.00 99000000.00 212850.00 22935.00 53793.00
127 100 44000.00 99000000.00 111262.50 22935.00 44000.00
85 100 111262.50 99000000.00 111262.50 22935.00 44000.00
34 110 61637.81 99000000.00 68805.00 22935.00 61637.81
61 110 68805.00 99000000.00 68805.00 22935.00 61637.81
110 115 6000000.00 99000000.00 7480000.00 22935.00 6000000.00
118 115 7480000.00 99000000.00 7480000.00 22935.00 6000000.00
28 120 22935.00 99000000.00 39224.06 22935.00 22935.00
37 120 39224.06 99000000.00 39224.06 22935.00 22935.00
121 125 99000000.00 99000000.00 99000000.00 22935.00 99000000.00



Built-in Windowing Functions : nth_value
Returns value of the nth row in a window

nth_value(salary,2) over (partition by dept_no order by salary desc)
from employee;
EMP_NO DEPT_NO SALARY NTH_VALUE
======= ======= ===================== =====================
105 000 212850.00 <null>
12 000 53793.00 53793.00
85 100 111262.50 <null>
127 100 44000.00 44000.00
61 110 68805.00 <null>
34 110 61637.81 61637.81
118 115 7480000.00 <null>
110 115 6000000.00 6000000.00
37 120 39224.06 <null>
36 120 33620.63 33620.63
28 120 22935.00 33620.63



Built-in Windowing Functions : last_value
Returns value of the last row in a window

last_value(salary) over (partition by dept_no order by salary asc)
from employee;
EMP_NO DEPT_NO SALARY LAST_VALUE
======= ======= ===================== =====================
12 000 53793.00 53793.00
105 000 212850.00 212850.00
127 100 44000.00 44000.00
85 100 111262.50 111262.50
34 110 61637.81 61637.81
61 110 68805.00 68805.00



Built-in Windowing Functions : lag and lead
Both LAG and LEAD functions have the same usage.

LAG (value [,offset[, default]]) OVER ([query_partition_clause] order_by_clause)
LEAD (value [,offset[, default]]) OVER ([query_partition_clause] order_by_clause)

• value - Can be a column or a built-in function, except for other analytic functions.
• offset - The number of rows preceeding/following the current row, from which the
data is to be retrieved. The default value is 1.
• default - The value returned if the offset is outside the scope of the window. The
default value is NULL.



Built-in Windowing Functions : lag
Returns value of row above

lag(salary) over (order by salary desc nulls last)
from employee;
EMP_NO DEPT_NO SALARY LAG
======= ======= ===================== =====================
121 125 99000000.00 <null>
118 115 7480000.00 99000000.00
110 115 6000000.00 7480000.00
134 123 390500.00 6000000.00
105 000 212850.00 390500.00
46 900 116100.00 212850.00
85 100 111262.50 116100.00
107 670 111262.50 111262.50
141 121 110000.00 111262.50

lag() only acts on a partition.



Built-in Windowing Functions : lead
Returns value of row below

lead(salary) over (order by salary desc nulls last)
from employee;
EMP_NO DEPT_NO SALARY LEAD
======= ======= ===================== =====================
121 125 99000000.00 7480000.00
118 115 7480000.00 6000000.00
110 115 6000000.00 390500.00
134 123 390500.00 212850.00
105 000 212850.00 116100.00
46 900 116100.00 111262.50
85 100 111262.50 111262.50
107 670 111262.50 110000.00
141 121 110000.00 105900.00
28 120 22935.00 <null>

lead() only acts on a partition.



Cumulative aggregates
The current row window is now restricted to all rows equal to or preceding the current
row within the current partition
• Find the total sales per quarter, and cumulative sales in quarter order

select order_year, order_quarter, sum(total_value) as q_sales,
sum(sum(total_value)) over (partition by order_year order by order_quarter)
as cum_sales_year
from v_sales_q
group by order_year, order_quarter;
ORDER_YEAR ORDER_QUARTER Q_SALES CUM_SALES_YEAR
========== ============= ===================== =====================
1992 3 2985.00 2985.00
1992 4 72000.00 74985.00
1993 1 648.00 648.00
1993 2 20000.00 20648.00
1993 3 1009643.04 1030291.04
1993 4 699724.16 1730015.20
1994 1 440590.83 440590.83



Performance
List orders, quantity ordered and cumulative quantity ordered by day

ORDER_DATE PO_NUMBER QTY_ORDERED QTY_CUMUL_DAY
=========== ========= ============ =====================
1991-03-04 V91E0210 10 10
1992-07-26 V92J1003 15 15
1992-10-15 V92E0340 7 7
1992-10-15 V92F3004 3 10
1993-02-03 V9333005 2 2
1993-03-22 V93C0120 1 1
1993-04-27 V9333006 5 5
1993-08-01 V93H3009 3 3
1993-08-09 V9324200 1000 1000
1993-08-09 V93C0990 40 1040



Performance
Without window function

SELECT ORDER_DATE, CUST_NO,QTY_ORDERED,
(SELECT SUM(QTY_ORDERED)
FROM SALES AS Si
WHERE Si.ORDER_DATE = S.ORDER_DATE
AND Si.CUST_NO <= S.CUST_NO)
AS QTY_CUMUL_DAY
FROM SALES AS S
ORDER BY S.ORDER_DATE, S.CUST_NO

PLAN (SI INDEX (RDB$FOREIGN25))
PLAN SORT (S NATURAL)
SALES 591 indexed reads
SALES 33 non indexed reads



Performance
With window function

SELECT ORDER_DATE, PO_NUMBER,QTY_ORDERED,
SUM(QTY_ORDERED)
OVER (PARTITION BY ORDER_DATE
ORDER BY PO_NUMBER)
AS QTY_CUMUL_DAY
FROM SALES
ORDER BY ORDER_DATE, PO_NUMBER

PLAN SORT (SALES NATURAL)



Performance
And you can extend it nearly without cost

SELECT ORDER_DATE, PO_NUMBER,QTY_ORDERED,
SUM(QTY_ORDERED)
OVER (PARTITION BY ORDER_DATE
ORDER BY PO_NUMBER)
AS QTY_CUMUL_DAY,
SUM(QTY_ORDERED)
OVER (PARTITION BY EXTRACT(YEAR FROM ORDER_DATE),EXTRACT(MONTH FROM ORDER_DATE)
ORDER BY ORDER_DATE, PO_NUMBER)
AS QTY_CUMUL_MONTH,
SUM(QTY_ORDERED)
OVER (PARTITION BY EXTRACT(YEAR FROM ORDER_DATE)
ORDER BY ORDER_DATE, PO_NUMBER)
AS QTY_CUMUL_YEAR
FROM SALES
ORDER BY ORDER_DATE, PO_NUMBER



PLAN SORT (SALES NATURAL)

ORDER_DATE PO_NUMBER QTY_ORDERED QTY_CUMUL_DAY QTY_CUMUL_MONTH QTY_CUMUL_YEAR
=========== ========= =========== ============= =============== ==============
1991-03-04 V91E0210 10 10 10 10
1992-07-26 V92J1003 15 15 15 15
1992-10-15 V92E0340 7 7 7 22
1992-10-15 V92F3004 3 10 10 25
1993-02-03 V9333005 2 2 2 2
1993-03-22 V93C0120 1 1 1 3
1993-04-27 V9333006 5 5 5 8
1993-08-01 V93H3009 3 3 3 11
1993-08-09 V9324200 1000 1000 1003 1011
1993-08-09 V93C0990 40 1040 1043 1051
1993-08-16 V9324320 1 1 1044 1052
1993-08-20 V93J3100 16 16 1060 1068
1993-08-27 V93F3088 10 10 1070 1078



Thank you !


Firebird 3 Windows Functions

Recommended

More Related Content

Similar to Firebird 3 Windows Functions (20)

More from Mind The Firebird (20)

Recently uploaded (20)

Firebird 3 Windows Functions