completed till Framing window function

Author: ptyadana

Complete SQL & Databases - ZTM/06_Advanced_SQL.sql

@@ -12,6 +12,7 @@ JOIN departments d ON d.dept_no = de.dept_no
 GROUP BY d.dept_no
 ORDER BY 1;
 
+/*------------------------------------------------------------------------------------------------------------*/
 
 /************ 2) HAVING Keyword *************/
 /*
@@ -45,6 +46,7 @@ GROUP BY d.dept_name
 HAVING COUNT(e.emp_no) > 25000
 ORDER BY 1;
 
+/*------------------------------------------------------------------------------------------------------------*/
 
 /********** 3) Ordering Group Data **********/
 SELECT d.dept_name AS "Department Name" ,COUNT(e.emp_no) AS "Number Of Employee"
@@ -56,13 +58,17 @@ HAVING COUNT(e.emp_no) > 25000
 ORDER BY 2 DESC;
 
 
-/********* 4) Group by Mental Model ***********/
+/********* 4) GROUP BY Mental Model ***********/
+
 /* What are the 8 employees who got the most salary bumps? */
 -- SELECT e.emp_no, CONCAT(e.first_name, e.last_name) AS "Name", s.salary, s.from_date, s.to_date
 SELECT emp_no, MAX(from_date)
 FROM salaries
 GROUP BY emp_no;
 
+
+/*------------------------------------------------------------------------------------------------------------*/
+
 /*********** 5) GROUPING SETS**********/
 
 /******* UNION / UNION ALL *********/
@@ -81,20 +87,25 @@ GROUP BY emp_no;
 */
 SELECT NULL AS "prod_id", sum(ol.quantity)
 FROM orderlines AS ol
+
 UNION
+
 SELECT prod_id AS "prod_id", sum(ol.quantity)
 FROM orderlines AS ol
 GROUP BY prod_id
 ORDER BY prod_id DESC;
 
+/*------------------------------------------------------------------------------------------------------------*/
+
 /*********** GROUPING SETS ***********/
 /* 
 	A Subclause of GROUP BY that allows you to define multiple grouping
 	It is very useful when we want to combine multiple grouping 
 */
 
--- same result as using above UNION code
-SELECT NULL AS "prod_id", sum(ol.quantity)
+-- same result as using above UNION code, but in same query
+-- here we are combining Two Sets (one for getting Total, one for per each product)
+SELECT prod_id, sum(ol.quantity)
 FROM orderlines AS ol
 GROUP BY
 	GROUPING SETS(
@@ -103,13 +114,213 @@ GROUP BY
 	)
 ORDER BY prod_id DESC;
 
+
 /* we can add in multiple groups as we need */
-SELECT NULL AS "prod_id", orderlineid, sum(ol.quantity)
+SELECT prod_id, orderlineid, sum(ol.quantity)
 FROM orderlines AS ol
 GROUP BY
 	GROUPING SETS(
 		(),
 		(prod_id),
 		(orderlineid)
 	)
-ORDER BY prod_id DESC, orderlineid DESC;
+ORDER BY prod_id DESC, orderlineid DESC;
+
+/*------------------------------------------------------------------------------------------------------------*/
+
+/************ GROUPING SETS for info from High Level to Details Level ***********/
+
+SELECT
+	EXTRACT(YEAR FROM orderdate) AS "YEAR",
+	EXTRACT(MONTH FROM orderdate) AS "MONTH",
+	EXTRACT(DAY FROM orderdate) AS "DAY",
+	SUM(quantity)AS "TOTAL QUANTITY"
+FROM orderlines
+GROUP BY 
+	GROUPING SETS(
+		(EXTRACT(YEAR FROM orderdate)), 		-- yearly
+		(EXTRACT(MONTH FROM orderdate)), 		-- monthly
+		(EXTRACT(DAY FROM orderdate)),			-- daily
+		(
+			EXTRACT(YEAR FROM orderdate),		-- month and year
+			EXTRACT(MONTH FROM orderdate)
+		),
+		(
+			EXTRACT(MONTH FROM orderdate),		-- month and day
+			EXTRACT(DAY FROM orderdate)
+		),
+		(
+			EXTRACT(YEAR FROM orderdate),		-- year, month and day
+			EXTRACT(MONTH FROM orderdate),
+			EXTRACT(DAY FROM orderdate)
+		),
+		()										-- nothing in particular (TOTAL AMOUNT)
+	)
+ORDER BY 1,2,3;
+
+/*------------------------------------------------------------------------------------------------------------*/
+
+/************			6) ROLLUP			***************/
+
+/* roll up can provide a very similar result as above using grouping sets, but with less code */
+SELECT
+	EXTRACT(YEAR FROM orderdate) AS "YEAR",
+	EXTRACT(MONTH FROM orderdate) AS "MONTH",
+	EXTRACT(DAY FROM orderdate) AS "DAY",
+	SUM(quantity)AS "TOTAL QUANTITY"
+FROM orderlines
+GROUP BY 
+	ROLLUP(
+		EXTRACT(YEAR FROM orderdate),
+		EXTRACT(MONTH FROM orderdate),
+		EXTRACT(DAY FROM orderdate)
+	)
+ORDER BY 1,2,3;
+
+/*------------------------------------------------------------------------------------------------------------*/
+
+/******************** 8/9) WINDOW Functions ******************/
+/*
+	Window functions CREATE a NEW COLUMN based on functions performed on a SUBSET or "WINDOW" of the data.
+	
+	window_function(agr1, agr2) OVER(
+		[PARTITION BY partition_expression]
+		[ORDER BY sort_expression [ASC | DESC] [NULLS {FIRST | LAST}]]
+	)
+*/
+
+-- Here we can see in the result that max salary is 158,220. Because query returns all data, then LIMIT say cut it off for 100 rows only. 
+-- That's why OVER() is calculated on the window or subset of data (in this case the entire data were returned).
+SELECT *,
+	MAX(salary) OVER()
+FROM salaries
+LIMIT 100;
+
+-- in this case, the maximum salary is 69,999. Because of WHERE conditions, the data were filtered out.
+-- and OVER() is using on that subset or window of the returned data (in this case the results of WHERE filtered data).
+SELECT 
+	*,
+	MAX(salary) OVER() 
+FROM salaries
+WHERE salary < 70000
+ORDER BY salary DESC;
+
+
+/******************** 10) PARTITON BY ******************/
+/*
+	Divide Rows into Groups to apply the function against (Optional)
+*/
+
+/* Employee salary compairing average salary of departments */
+SELECT 
+	s.emp_no, s.salary,d.dept_name,
+	AVG(s.salary)
+	OVER(
+		PARTITION BY(d.dept_name)
+	)
+FROM salaries s
+JOIN dept_emp de ON s.emp_no = de.emp_no
+JOIN departments d ON d.dept_no = de.dept_no;
+
+
+/******************** 11) ORDER BY ******************/
+/*
+	ORDER BY changes the FRAME of the window
+	It tells SQL to take into account of everything before up until to this point (becoming Cumulative)
+*/
+-- against the window of entire data
+SELECT emp_no,
+	COUNT(salary) OVER()
+FROM salaries;
+
+-- using PARTION BY
+-- Counting salary by each unique emp_no partion
+SELECT emp_no,
+	COUNT(salary) OVER(
+		PARTITION BY(emp_no)
+	)
+FROM salaries;
+
+
+-- using ORDER BY
+-- Count number are becoming Cumulative
+SELECT emp_no,
+	COUNT(salary) OVER(
+		ORDER BY emp_no
+	)
+FROM salaries;
+
+
+/********************************* FRAME Clause *******************************************************/
+/*
+	When using Frame clause in a window function, we can create a SUB-RANGE or FRAME
+	
+	For example: when we use ORDER BY, we look at the PARTATIONED data in a different len (FRAME).
+	
+	NOTE: 
+	Without ORDER BY, by default the framing is usually ALL PARTITION ROWs (Entire Window)
+	With ORDER BY, by default the framing is usually everything before the CURRENT ROW AND the CURRENT ROW (Cumulatively)
+	
+	|----------------------------------------------------------------------------------------------------
+	|			Keys						| 			Meaning 										|
+	|---------------------------------------|-----------------------------------------------------------|
+	|	ROWS or RANGE						|	Whether you want to use a RANGE or ROWS as a FRAME		|
+	|	PRECEDING							|	Rows Before the CURRENT ONE								|
+	|	FOLLOWING							|	Rows After the CURRENT ONE								|
+	| 	UNBOUNDED PRECEDING or FOLLOWING	|	Returns All Before and After							|
+	|	CURRENT ROW							|	Your Current Row										|
+	-----------------------------------------------------------------------------------------------------
+*/
+
+-- In this case, we can see that every salary is unique. Because we are using ORDER BY, each row is cumulativly
+-- counted within the partition of each employee's window. As a result, it like like 1, (1+1 becomes 2), etc.
+SELECT emp_no,
+	salary,
+	COUNT(salary) OVER(
+		PARTITION BY emp_no
+		ORDER BY salary
+	)
+FROM Salaries;
+
+
+-- This one returns the same results as using PARTION BY only. The reason is we are looking at the data through the lends of Range.
+-- For range using unbounded precedning and following, we are comparing against data with the entire data within that Partition.
+SELECT emp_no,
+	salary,
+	COUNT(salary) OVER(
+		PARTITION BY emp_no
+		ORDER BY salary
+		RANGE BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING
+	)
+FROM Salaries;
+
+-- same reults as RANGE results
+SELECT emp_no,
+	salary,
+	COUNT(salary) OVER(
+		PARTITION BY emp_no
+		ORDER BY salary
+		ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING
+	)
+FROM Salaries;
+
+
+-- same like ORDER BY
+SELECT emp_no,
+	salary,
+	COUNT(salary) OVER(
+		PARTITION BY emp_no
+		ORDER BY salary
+		ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW
+	)
+FROM Salaries;
+
+/*------------------------------------------------------------------------------------------------------------*/
+
+
+
+
+
+
+
+