Windowing Functions - Little Rock Tech fest 2019

The following is intended to outline our general product direction. It is intended for information purposes
only, and may not be incorporated into any contract. It is not a commitment to deliver any material, code,
or functionality, and should not be relied upon in making purchasing decisions. The development,
release, timing, and pricing of any features or functionality described for Oracle’s products may change
and remains at the sole discretion of Oracle Corporation.
Statements in this presentation relating to Oracle’s future plans, expectations, beliefs, intentions and
prospects are “forward-looking statements” and are subject to material risks and uncertainties. A detailed
discussion of these factors and other risks that affect our business is contained in Oracle’s Securities and
Exchange Commission (SEC) filings, including our most recent reports on Form 10-K and Form 10-Q
under the heading “Risk Factors.” These filings are available on the SEC’s website or on Oracle’s website
at http://www.oracle.com/investor. All information in this presentation is current as of September 2019
and Oracle undertakes no duty to update any statement in light of new information or future events.
Safe Harbor
Copyright © 2019 Oracle and/or its affiliates.

SQL Windowing Functions
Dave Stokes
MySQL Community Manager
Oracle Corporation

Window Functions
Window functions allow access to data in
the records right before and after the
current record.
A window function defines a frame or
window of rows with a given length around
the current row, and performs a calculation
across the set of data in the window.
https://en.wikipedia.org/wiki/SQL_window_function

Windowing Functions
MySQL 8.0 added support for window
functions.
However we will start with what doing
calculations WITHOUT WFs looked like!

Sample Data – w/o WF
SQL > select a,b,c,d from w1;
+---+----+-----+------+
| a | b | c | d |
+---+----+-----+------+
| 1 | 10 | 100 | 1000 |
| 2 | 20 | 200 | 2000 |
| 3 | 30 | 300 | 3000 |
| 4 | 40 | 400 | 4000 |
+---+----+-----+------+
4 rows in set (0.0004 sec)

Try to add rows -> Opps!
select a,b,c,d, sum(a+b) as 'a&b' from w1;
+---+----+-----+------+-----+
| a | b | c | d | a&b |
+---+----+-----+------+-----+
| 1 | 10 | 100 | 1000 | 110 |
+---+----+-----+------+-----+
1 row in set (0.0028 sec)
1+2+3+4+10+20+30+40 = 110

Tell server how to ‘group’ data -> Better!
select a,b,c,d, sum(a+b) as 'a&b'
from w1
GROUP BY a;
+---+----+-----+------+-----+
| a | b | c | d | a&b |
+---+----+-----+------+-----+
| 1 | 10 | 100 | 1000 | 11 |
| 2 | 20 | 200 | 2000 | 22 |
| 3 | 30 | 300 | 3000 | 33 |
| 4 | 40 | 400 | 4000 | 44 |
+---+----+-----+------+-----+
4 rows in set (0.0019 sec)

We can sum columns but …
select a, sum(a), b, sum(b) from w1;
+---+--------+----+--------+
| a | sum(a) | b | sum(b) |
+---+--------+----+--------+
| 1 | 10 | 10 | 100 |
+---+--------+----+--------+
We can sum the first value
from a row and the sum of
all ‘a’ values. But does that
first value really tell us
anything?
1+2+3+4 = 10
10+20+30+40 = 100

New Data
SQL > select id, price, warehouse, vendor
from w2;
+----+-------+-----------+--------+
| id | price | warehouse | vendor |
+----+-------+-----------+--------+
| 1 | 1.99 | 1 | 1 |
| 2 | 10.50 | 1 | 2 |
| 3 | 0.99 | 2 | 2 |
| 4 | 1.10 | 1 | 2 |
+----+-------+-----------+--------+

Grouping and Rollup
select warehouse,
sum(price)
from w2
group by warehouse WITH ROLLUP;
+-----------+------------+
| warehouse | sum(price) |
+-----------+------------+
| 1 | 13.59 |
| 2 | 0.99 |
| NULL | 14.58 |
+-----------+------------+
We can group like items
together and even ‘roll up’
values for totals.
The NULL under the
warehouse column is the
ROLLUP or total of the
sum(price) -- And not easily
understood

And we can use different columns
SQL > select vendor,
sum(price)
from w2
group by vendor with rollup;
+--------+------------+
| vendor | sum(price) |
+--------+------------+
| 1 | 1.99 |
| 2 | 12.59 |
| NULL | 14.58 |
+--------+------------+

Non Windowing Aggregate Functions
• SUM
• COUNT
• MAX,MIN,
• STD, STDDEV,STDDEV_POP, STDDEV_SAMP
• VAR,VAR_POP, VARIANCE
• See https://dev.mysql.com/doc/refman/8.0/en/group-by-functions.html#function_sum

So Why Windowing Functions
A window function performs an aggregate-like operation
on a set of query rows. However, whereas an aggregate
operation groups query rows into a single result row, a
window function produces a result for each query row:
• The row for which function evaluation occurs is called
the current row.
• The query rows related to the current row over which
function evaluation occurs comprise the window for the
current row.

Warning!
Windowing Functions are difficult – you need
to practice with them to build understanding
and competence.
Do not panic if you struggle at first – they are
a learned skilled.

NULL
Null (or NULL) is a special marker used in Structured Query
Language to indicate that a data value does not exist in the
database. Introduced by the creator of the relational database model,
E. F. Codd, SQL Null serves to fulfil the requirement that all true
relational database management systems (RDBMS) support a
representation of "missing information and inapplicable information“
https://en.wikipedia.org/wiki/Null_(SQL)

New Keyword - OVER
SELECT year, country, product, profit,
SUM(profit) OVER() AS total_profit,
SUM(profit) OVER(PARTITION BY country) AS country_profit
FROM sales
ORDER BY country, year, product, profit;
+------+---------+------------+--------+--------------+----------------+
| year | country | product | profit | total_profit | country_profit |
+------+---------+------------+--------+--------------+----------------+
| 2000 | Finland | Computer | 1500 | 7535 | 1610 | 1610 = 1500+100+10 (Finland)
| 2000 | Finland | Phone | 100 | 7535 | 1610 |
| 2001 | Finland | Phone | 10 | 7535 | 1610 |
| 2000 | India | Calculator | 75 | 7535 | 1350 |
| 2000 | India | Calculator | 75 | 7535 | 1350 |
| 2000 | India | Computer | 1200 | 7535 | 1350 |
| 2000 | USA | Calculator | 75 | 7535 | 4575 |
| 2000 | USA | Computer | 1500 | 7535 | 4575 |
| 2001 | USA | Calculator | 50 | 7535 | 4575 |
| 2001 | USA | Computer | 1200 | 7535 | 4575 |
| 2001 | USA | Computer | 1500 | 7535 | 4575 |
| 2001 | USA | TV | 100 | 7535 | 4575 |
| 2001 | USA | TV | 150 | 7535 | 4575 |
+------+---------+------------+--------+--------------+----------------+
= 7535

New Keyword - OVER
SELECT year, country, product, profit,
SUM(profit) OVER() AS total_profit,
SUM(profit) OVER(PARTITION BY country) AS country_profit
FROM sales
ORDER BY country, year, product, profit;
The first OVER() clause is empty which treats the entire set of rows as a
partition (global).
The second OVER() clause partitions rows by country, producing a sum per
partition (per country). The function produces this sum for each partition
row (country).

Supported Non Aggregate Functions
• CUME_DIST()
• DENSE_RANK()
• FIRST_VALUE()
• LAG(), LEAD()
• LAST_VALUE()
• NTH_VALUE()
• NTILE()
• PERCENT_RANK()
• RANK()
• ROW_NUMBER()

ROW_NUMBER()
SQL > select
ROW_NUMBER() over () as 'row nbr', b
FROM w1;
+---------+----+
| row nbr | b |
+---------+----+
| 1 | 10 |
| 2 | 20 |
| 3 | 30 |
| 4 | 40 |
+---------+----+

Rank and Dense Rank
SELECT x, row_number() over (order by x) AS 'Row Nbr',
rank() over (order by x) AS 'Rank',
DENSE_RANK() over (order by x) as 'Dense Rank'
from w4;
+---+---------+------+------------+
| x | Row Nbr | Rank | Dense Rank |
+---+---------+------+------------+
| 0 | 1 | 1 | 1 |
| 0 | 2 | 1 | 1 |
| 2 | 3 | 3 | 2 |
| 3 | 4 | 4 | 3 |
| 3 | 5 | 4 | 3 |
| 4 | 6 | 6 | 4 |
+---+---------+------+------------+
+---+
| x |
+---+
| 0 |
| 0 |
| 2 |
| 3 |
| 3 |
| 4 |
+---+

Rank and Dense Rank – Messy repeat
SELECT x, row_number() over (order by x) AS 'Row Nbr',
rank() over (order by x) AS 'Rank',
DENSE_RANK() over (order by x) as 'Dense Rank'
from w4;
+---+---------+------+------------+
+---+---------+------+------------+
| 0 | 1 | 1 | 1 |
| 0 | 2 | 1 | 1 |
| 2 | 3 | 3 | 2 |
| 3 | 4 | 4 | 3 |
| 3 | 5 | 4 | 3 |
| 4 | 6 | 6 | 4 |
+---+---------+------+------------+
+---+
| x |
+---+
| 0 |
| 0 |
| 2 |
| 3 |
| 3 |
| 4 |
+---+

Rank and Dense Rank – Named Window
SELECT x, ROW_NUMBER() over w AS 'Row Nbr',
RANK() over w AS 'Rank',
DENSE_RANK() over w as 'Dense Rank'
from w4
WINDOW w as (order by x);
+---+---------+------+------------+
+---+---------+------+------------+
| 0 | 1 | 1 | 1 |
| 0 | 2 | 1 | 1 |
| 2 | 3 | 3 | 2 |
| 3 | 4 | 4 | 3 |
| 3 | 5 | 4 | 3 |
| 4 | 6 | 6 | 4 |
+---+---------+------+------------+
+---+
| x |
+---+
| 0 |
| 0 |
| 2 |
| 3 |
| 3 |
| 4 |
+---+

You Can Modify Window Clauses
SELECT DISTINCT year,
country,
FIRST_VALUE(year) OVER (w ORDER BY year ASC) AS first,
FIRST_VALUE(year) OVER (w ORDER BY year DESC) AS last
FROM sales
WINDOW w AS (PARTITION BY country);

What if you double up?
select date,
name,
first_value(date) over (w order by name) as first
from sales
window w as (order by date);
ERROR: 3583: Window '<unnamed window>' cannot inherit 'w'
since both contain an ORDER BY clause.

Another Data Set
select * from sales;
+------+------------+-------+
| name | date | sales |
+------+------------+-------+
| Dave | 2019-01-01 | 125 |
| Jack | 2019-01-15 | 86 |
| Lucy | 2019-01-13 | 140 |
| Dave | 2019-03-03 | 100 |
| Lucy | 2019-04-01 | 200 |
| Dave | 2019-09-01 | 100 |
| Jack | 2019-09-15 | 95 |
| Lucy | 2019-09-13 | 140 |
| Jack | 2019-06-03 | 100 |
| Lucy | 2019-05-01 | 200 |
+------+------------+-------+
It is often common to want to know the
sales by name, calendar dates, ranges
of prices, and etcetera.

Another Data Set
+------+------------+-------+
+------+------------+-------+
| Dave | 2019-01-01 | 125 |
| Jack | 2019-01-15 | 86 |
| Lucy | 2019-01-13 | 140 |
| Dave | 2019-03-03 | 100 |
| Lucy | 2019-04-01 | 200 |
| Dave | 2019-09-01 | 100 |
| Jack | 2019-09-15 | 95 |
| Lucy | 2019-09-13 | 140 |
| Jack | 2019-06-03 | 100 |
| Lucy | 2019-05-01 | 200 |
+------+------------+-------+
SELECT name,
sum(sales)
from sales
group by name;
+------+------------+
| name | sum(sales) |
+------+------------+
| Dave | 325 |
| Jack | 281 |
| Lucy | 680 |
+------+------------+

Another Data Set
+------+------------+-------+
+------+------------+-------+
| Dave | 2019-01-01 | 125 |
| Jack | 2019-01-15 | 86 |
| Lucy | 2019-01-13 | 140 |
| Dave | 2019-03-03 | 100 |
| Lucy | 2019-04-01 | 200 |
| Dave | 2019-09-01 | 100 |
| Jack | 2019-09-15 | 95 |
| Lucy | 2019-09-13 | 140 |
| Jack | 2019-06-03 | 100 |
| Lucy | 2019-05-01 | 200 |
+------+------------+-------+
select name, date, sales,
SUM(sales) OVER (partition by name) as 'sum'
FROM sales;
+------+------------+-------+-----+
| name | date | sales | sum |
+------+------------+-------+-----+
| Dave | 2019-01-01 | 125 | 325 |
| Dave | 2019-03-03 | 100 | 325 |
| Dave | 2019-09-01 | 100 | 325 |
| Jack | 2019-01-15 | 86 | 281 |
| Jack | 2019-09-15 | 95 | 281 |
| Jack | 2019-06-03 | 100 | 281 |
| Lucy | 2019-01-13 | 140 | 680 |
| Lucy | 2019-04-01 | 200 | 680 |
| Lucy | 2019-09-13 | 140 | 680 |
| Lucy | 2019-05-01 | 200 | 680 |
+------+------------+-------+-----+

Monthly results
SELECT name, MONTHNAME(date), sales,
sum(sales) OVER (PARTITION BY MONTH(date)) as 'my sum'
FROM sales;
+------+-----------------+-------+--------+
| name | MONTHNAME(date) | sales | my sum |
+------+-----------------+-------+--------+
| Dave | January | 125 | 351 |
| Jack | January | 86 | 351 |
| Lucy | January | 140 | 351 |
| Dave | March | 100 | 100 |
| Lucy | April | 200 | 200 |
| Lucy | May | 200 | 200 |
| Jack | June | 100 | 100 |
| Dave | September | 100 | 335 |
| Jack | September | 95 | 335 |
| Lucy | September | 140 | 335 |
+------+-----------------+-------+--------+

Cumulative Sales by name
SELECT name, sales, date,
SUM(sales) OVER (PARTITION by name ORDER BY date) as cum_sales
FROM sales;
+------+-------+------------+-----------+
| name | sales | date | cum_sales |
+------+-------+------------+-----------+
| Dave | 125 | 2019-01-01 | 125 |
| Dave | 100 | 2019-03-03 | 225 |
| Dave | 100 | 2019-09-01 | 325 |
| Jack | 86 | 2019-01-15 | 86 |
| Jack | 100 | 2019-06-03 | 186 |
| Jack | 95 | 2019-09-15 | 281 |
| Lucy | 140 | 2019-01-13 | 140 |
| Lucy | 200 | 2019-04-01 | 340 |
| Lucy | 200 | 2019-05-01 | 540 |
| Lucy | 140 | 2019-09-13 | 680 |
+------+-------+------------+-----------+

Current, N, and Unbounded Rows

Cumulative Sales by sale
SELECT name, sales, date,
sum(sales) over (ORDER BY date ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
as cum_sales
FROM sales;
+------+-------+------------+-----------+
| name | sales | date | cum_sales |
+------+-------+------------+-----------+
| Dave | 125 | 2019-01-01 | 125 | 125
| Lucy | 140 | 2019-01-13 | 265 | 125+140 = 265
| Jack | 86 | 2019-01-15 | 351 | 265 + 86 = 351
| Dave | 100 | 2019-03-03 | 451 |
| Lucy | 200 | 2019-04-01 | 651 |
| Lucy | 200 | 2019-05-01 | 851 |
| Jack | 100 | 2019-06-03 | 951 |
| Dave | 100 | 2019-09-01 | 1051 |
| Lucy | 140 | 2019-09-13 | 1191 |
| Jack | 95 | 2019-09-15 | 1286 |
+------+-------+------------+-----------+

Average Sales
SELECT MONTH(date), SUM(sales),
AVG(SUM(sales)) OVER (ORDER BY MONTH(date) RANGE BETWEEN 1 PRECEDING and 1 FOLLOWING)
AS slidingAvg from sales group by MONTH(date);
+-------------+------------+------------+
| MONTH(date) | SUM(sales) | slidingAvg |
+-------------+------------+------------+
| 1 | 351 | 351.0000 | 351
| 3 | 100 | 150.0000 | 150 = (351+100) / 3
| 4 | 200 | 166.6667 | 166.6667 = (100+200+200) / 3
| 5 | 200 | 166.6667 |
| 6 | 100 | 150.0000 |
| 9 | 335 | 335.0000 |
+-------------+------------+------------+

Another Data Set
+------+------------+-------+
+------+------------+-------+
| Dave | 2019-01-01 | 125 |
| Jack | 2019-01-15 | 86 |
| Lucy | 2019-01-13 | 140 |
| Dave | 2019-03-03 | 100 |
| Lucy | 2019-04-01 | 200 |
| Dave | 2019-09-01 | 100 |
| Jack | 2019-09-15 | 95 |
| Lucy | 2019-09-13 | 140 |
| Jack | 2019-06-03 | 100 |
| Lucy | 2019-05-01 | 200 |
+------+------------+-------+
select name, date, sales,
SUM(sales) OVER w as 'sum'
FROM sales
WINDOW w AS (partition by name order by date);
+------+------------+-------+-----+
| name | date | sales | sum |
+------+------------+-------+-----+
| Dave | 2019-01-01 | 125 | 125 |
| Dave | 2019-03-03 | 100 | 225 |
| Dave | 2019-09-01 | 100 | 325 |
| Jack | 2019-01-15 | 86 | 86 |
| Jack | 2019-06-03 | 100 | 186 |
| Jack | 2019-09-15 | 95 | 281 |
| Lucy | 2019-01-13 | 140 | 140 |
| Lucy | 2019-04-01 | 200 | 340 |
| Lucy | 2019-05-01 | 200 | 540 |
| Lucy | 2019-09-13 | 140 | 680 |
+------+------------+-------+-----+

Syntax
over_clause:
{OVER (window_spec) | OVER window_name}
window_spec:
[window_name] [partition_clause] [order_clause]
[frame_clause]
partition_clause:
PARTITION BY expr [, expr] ...
order_clause:
ORDER BY expr [ASC|DESC] [, expr [ASC|DESC]] ...

Syntax
frame_clause: frame_units frame_extent
frame_units: {ROWS | RANGE}
frame_extent:
{frame_start | frame_between}
frame_between:
BETWEEN frame_start AND frame_end frame_start, frame_end:
{ CURRENT ROW | UNBOUNDED PRECEDING | UNBOUNDED FOLLOWING |
expr PRECEDING | expr FOLLOWING }

Lets work a simple average w/ default
select a, count(a) over w as 'count a',
b, avg(b) over w as 'avg b' from w1
window w as (order by a) ;
+---+---------+----+---------+
| a | count a | b | avg b |
+---+---------+----+---------+
| 1 | 1 | 10 | 10.0000 |
| 2 | 2 | 20 | 15.0000 | 15 = (10+20)/2
| 3 | 3 | 30 | 20.0000 | 30 = (10+20+30)/3
| 4 | 4 | 40 | 25.0000 | 25 = 100/4
+---+---------+----+---------+ Default is UNBOUNDED
PRECEDING to CURRENT

ROWS BETWEEN
window w as
(ROWS BETWEEN 1 PRECEDING and CURRENT ROW) ;
+---+---------+----+---------+
+---+---------+----+---------+
| 1 | 1 | 10 | 10.0000 |
| 2 | 2 | 20 | 15.0000 |
| 3 | 2 | 30 | 25.0000 | 25 = (20+30) / 2
| 4 | 2 | 40 | 35.0000 | 35 = (30+40) / 2

ROWS BETWEEN
window w as
(ROWS BETWEEN UNBOUNDED PRECEDING and CURRENT ROW) ;
+---+---------+----+---------+
+---+---------+----+---------+
| 1 | 1 | 10 | 10.0000 |
| 2 | 2 | 20 | 15.0000 |
| 3 | 3 | 30 | 20.0000 |
| 4 | 4 | 40 | 25.0000 |

ROWS BETWEEN
window w as (PARTITION BY a >= 3) ;
+---+---------+----+---------+
+---+---------+----+---------+
| 1 | 2 | 10 | 15.0000 | 15 = (10+20)/2
| 2 | 2 | 20 | 15.0000 |
| 3 | 2 | 30 | 35.0000 | Start of PARTITION
| 4 | 2 | 40 | 35.0000 | 35 = (30+40)/2

LAG – values from preceding (x) rows
select a, lag(a,1) over w as 'lag(1)',
lag(a,2) over w as 'lag(2)' from w1
window w as (order by a);
+---+--------+--------+
| a | lag(1) | lag(2) |
+---+--------+--------+
| 1 | NULL | NULL |
| 2 | 1 | NULL |
| 3 | 2 | 1 |
| 4 | 3 | 2 |
+---+--------+--------+

LEAD – values from preceding (x) rows
select a, lead(a,1) over w as 'lead(1)',
lead(a,2) over w as 'lead(2)' from w1
window w as (order by a);
+---+---------+---------+
| a | lead(1) | lead(2) |
+---+---------+---------+
| 1 | 2 | 3 |
| 2 | 3 | 4 |
| 3 | 4 | NULL |
| 4 | NULL | NULL |

First and Last Values
select a, FIRST_VALUE(a) over w as 'first value',
LAST_VALUE(a) over w as 'last value'
from w1 window w as (order by a);
+---+-------------+------------+
| a | first value | last value |
+---+-------------+------------+
| 1 | 1 | 1 |
| 2 | 1 | 2 |
| 3 | 1 | 3 |
| 4 | 1 | 4 |

N Values
select a,
NTH_VALUE(a,2) over w as 'n2',
NTH_VALUE(a,3) over w as 'n3' from w1 window w as
(order by a);
+---+------+------+
| a | n2 | n3 |
+---+------+------+
| 1 | NULL | NULL | No 2nd or 3rd row read so no values
| 2 | 2 | NULL | No 3rd row read so no values
| 3 | 2 | 3 |
| 4 | 2 | 3 |

Another Example
SELECT * FROM staff ORDER by id;
+----+-------+-----------+---------+
| id | name | dept | salary |
+----+-------+-----------+---------+
| 1 | Andy | Shipping | 5400.00 |
| 2 | Betty | Marketing | 6300.00 |
| 3 | Tracy | Shipping | 4800.00 |
| 4 | Mike | Marketing | 7100.00 |
| 5 | Sandy | Sales | 5400.00 |
| 6 | James | Shipping | 6000.00 |
| 7 | Carol | Sales | 4600.00 |

Another Example
SELECT COUNT(*), SUM(salary) FROM staff;
+----------+-------------+
| COUNT(*) | SUM(salary) |
+----------+-------------+
| 7 | 39600.00 |
+----------+-------------+

Another Example
SELECT COUNT(*) AS 'nbr' ,
SUM(salary) as 'tot salary',
ROUND(avg(salary),0) AS avg
FROM staff
GROUP BY dept ORDER BY dept;
+-----+------------+------+
| nbr | tot salary | avg |
+-----+------------+------+
| 2 | 13400.00 | 6700 |
| 2 | 10000.00 | 5000 |
| 3 | 16200.00 | 5400 |

Another Example
SELECT dept, COUNT(*) AS 'nbr' ,
SUM(salary) as 'tot salary',
ROUND(avg(salary),0) AS avg
FROM staff
GROUP BY dept WITH ROLLUP order by dept;
+-----------+-----+------------+------+
| dept | nbr | tot salary | avg |
+-----------+-----+------------+------+
| NULL | 7 | 39600.00 | 5657 |
| Marketing | 2 | 13400.00 | 6700 |
| Sales | 2 | 10000.00 | 5000 |
| Shipping | 3 | 16200.00 | 5400 |

Calculate Rank
SELECT name, salary,
RANK() OVER s AS 'Rank' ,
DENSE_RANK() OVER s AS ‘Dense Rank'
FROM staff
WINDOW s as (ORDER BY salary DESC) ;
+-------+---------+------+-------------+
| name | salary | Rank | Dense Rank |
+-------+---------+------+-------------+
| Mike | 7100.00 | 1 | 1 |
| Betty | 6300.00 | 2 | 2 |
| James | 6000.00 | 3 | 3 |
| Sandy | 5400.00 | 4 | 4 |
| Andy | 5400.00 | 4 | 4 |
| Tracy | 4800.00 | 6 | 5 |
| Carol | 4600.00 | 7 | 6 |

Percentage of Total Salary
ROUND(salary / SUM(salary) OVER() * 100,2) AS '%'
FROM staff ORDER BY salary DESC;
+-------+---------+-------+
| name | salary | % |
+-------+---------+-------+
| Mike | 7100.00 | 17.93 |
| Betty | 6300.00 | 15.91 |
| James | 6000.00 | 15.15 |
| Andy | 5400.00 | 13.64 |
| Sandy | 5400.00 | 13.64 |
| Tracy | 4800.00 | 12.12 |
| Carol | 4600.00 | 11.62 |

Difference to AVG Salary
ROUND(AVG(salary) OVER (),2) AS 'avg',
ROUND(salary - AVG(salary) OVER (),2) AS 'Diff to AVG'
+-------+---------+---------+-------------+
| name | salary | avg | Diff to AVG |
+-------+---------+---------+-------------+
| Mike | 7100.00 | 5657.14 | 1442.86 |
| Betty | 6300.00 | 5657.14 | 642.86 |
| James | 6000.00 | 5657.14 | 342.86 |
| Andy | 5400.00 | 5657.14 | -257.14 |
| Sandy | 5400.00 | 5657.14 | -257.14 |
| Tracy | 4800.00 | 5657.14 | -857.14 |
| Carol | 4600.00 | 5657.14 | -1057.14 |

Calculate differences in salary
salary - LEAD(salary,1) OVER (ORDER BY salary DESC) as 'diff next'
+-------+---------+-----------+
| name | salary | diff next |
+-------+---------+-----------+
| Mike | 7100.00 | 800.00 |
| Betty | 6300.00 | 300.00 |
| James | 6000.00 | 600.00 |
| Andy | 5400.00 | 0.00 |
| Sandy | 5400.00 | 600.00 |
| Tracy | 4800.00 | 200.00 |
| Carol | 4600.00 | NULL |

salary - LAST_VALUE(salary) OVER w AS 'above',
ROUND((salary - LAST_VALUE(salary) OVER w) / LAST_VALUE(salary)
OVER w * 100) AS 'pct above'
FROM staff
WINDOW w as (ORDER by salary DESC ROWS BETWEEN UNBOUNDED PRECEDING AND
UNBOUNDED FOLLOWING) ORDER BY salary DESC;
+-------+---------+---------+-----------+
| name | salary | above | pct above |
+-------+---------+---------+-----------+
| Mike | 7100.00 | 2500.00 | 54 |
| Betty | 6300.00 | 1700.00 | 37 |
| James | 6000.00 | 1400.00 | 30 |
| Sandy | 5400.00 | 800.00 | 17 |
| Andy | 5400.00 | 800.00 | 17 |
| Tracy | 4800.00 | 200.00 | 4 |
| Carol | 4600.00 | 0.00 | 0 |

SELECT name, dept, salary,
round(AVG(salary) OVER z,2) AS 'Avg',
round(salary - AVG(salary) OVER z,2) as 'diff avg'
FROM staff WINDOW z AS (PARTITION BY dept)
ORDER BY dept, salary DESC;
+-------+-----------+---------+---------+----------+
| name | dept | salary | Avg | diff avg |
+-------+-----------+---------+---------+----------+
| Mike | Marketing | 7100.00 | 6700.00 | 400.00 |
| Betty | Marketing | 6300.00 | 6700.00 | -400.00 |
| Sandy | Sales | 5400.00 | 5000.00 | 400.00 |
| Carol | Sales | 4600.00 | 5000.00 | -400.00 |
| James | Shipping | 6000.00 | 5400.00 | 600.00 |
| Andy | Shipping | 5400.00 | 5400.00 | 0.00 |
| Tracy | Shipping | 4800.00 | 5400.00 | -600.00 |

Company and Departmental Ranks
SELECT name, dept, salary,
RANK() OVER s AS dept_rank,
RANK() OVER (ORDER BY salary DESC) AS Company_rank
FROM staff
WINDOW s as (PARTITION BY dept ORDER BY salary DESC)
ORDER BY dept, salary DESC;
+-------+-----------+---------+-----------+--------------+
| name | dept | salary | dept_rank | Company_rank |
+-------+-----------+---------+-----------+--------------+
| Mike | Marketing | 7100.00 | 1 | 1 |
| Betty | Marketing | 6300.00 | 2 | 2 |
| Sandy | Sales | 5400.00 | 1 | 4 |
| Carol | Sales | 4600.00 | 2 | 7 |
| James | Shipping | 6000.00 | 1 | 3 |
| Andy | Shipping | 5400.00 | 2 | 4 |
| Tracy | Shipping | 4800.00 | 3 | 6 |

Good Place to Stop .. for now.

Excellent Tutorial on Windowing Functions
https://momjian.us/main/writings/pgsql/window.pdf

MySQL Windowing Functions
https://dev.mysql.com/doc/refman/8.0/en/window-functions.html

Please buy my Book!

Q/A
Slides at Slideshare.net/davidmstokes
Email at David.Stokes @ Oracle.com
Blog at https://elephantdolphin.blogspot.com/

Windowing Functions - Little Rock Tech fest 2019

More Related Content

What's hot

Similar to Windowing Functions - Little Rock Tech fest 2019

More from Dave Stokes

Recently uploaded

Windowing Functions - Little Rock Tech fest 2019