filtering, join and aggregrations

Author: ptyadana

Advanced SQL for Data Scientists/03.Filtering_Join_and_Aggregrationsql.sql

@@ -0,0 +1,262 @@
+/********** Filterig, Join and Aggregration ************/
+
+/* we want to know person's salary comparing to his/her department average salary */
+SELECT
+	s.last_name,s.salary,s.department,
+	(SELECT ROUND(AVG(salary),2)
+	 FROM staff s2
+	 WHERE s2.department = s.department) AS department_average_salary
+FROM staff s;
+
+
+/* how many people are earning above/below the average salary of his/her department ? */
+CREATE VIEW vw_salary_comparision_by_department
+AS
+	SELECT 
+	s.department,
+	(
+		s.salary > (SELECT ROUND(AVG(s2.salary),2)
+					 FROM staff s2
+					 WHERE s2.department = s.department)
+	)AS is_higher_than_dept_avg_salary
+	FROM staff s
+	ORDER BY s.department;
+	
+	
+SELECT * FROM vw_salary_comparision_by_department;
+
+SELECT department, is_higher_than_dept_avg_salary, COUNT(*) AS total_employees
+FROM vw_salary_comparision_by_department
+GROUP BY 1,2;
+
+
+----------------------------------------------------------------------------------------------
+
+/* Assume that people who earn at latest 100,000 salary is Executive.
+We want to know the average salary for executives for each department.
+
+Data Interpretation: it seem like Sports department has the highest average salary for Executives
+where Movie department has the lowest.*/
+SELECT department, ROUND(AVG(salary),2) AS average_salary
+FROM staff
+WHERE salary >= 100000
+GROUP BY department
+ORDER BY 2 DESC;
+
+
+/* who earn the most in the company? 
+It seems like Stanley Grocery earns the most.
+*/
+SELECT last_name, department, salary
+FROM staff
+WHERE salary = (
+	SELECT MAX(s2.salary)
+	FROM staff s2
+);
+
+
+
+/* who earn the most in his/her own department */
+SELECT s.department, s.last_name, s.salary
+FROM staff s
+WHERE s.salary = (
+	SELECT MAX(s2.salary)
+	FROM staff s2
+	WHERE s2.department = s.department
+)
+ORDER BY 1;
+	
+----------------------------------------------------------------------------------------------
+
+SELECT * FROM company_divisions;
+
+/* full details info of employees with company division
+Based on the results, we see that there are only 953 rows returns. We know that there are 1000 staffs.
+*/
+SELECT s.last_name, s.department, cd.company_division
+FROM staff s
+JOIN company_divisions cd
+	ON cd.department = s.department;
+	
+	
+/* now all 1000 staffs are returned, but some 47 people have missing company - division.*/
+SELECT s.last_name, s.department, cd.company_division
+FROM staff s
+LEFT JOIN company_divisions cd
+	ON cd.department = s.department;
+	
+	
+/* who are those people with missing company division? 
+Data Interpretation: it seems like all staffs from "books" department have missing company division.
+We may want to inform our IT team to add Books department in corresponding company division.
+*/
+SELECT s.last_name, s.department, cd.company_division
+FROM staff s
+LEFT JOIN company_divisions cd
+	ON cd.department = s.department
+WHERE company_division IS NULL;
+	
+
+----------------------------------------------------------------------------------------------
+
+CREATE VIEW vw_staff_div_reg AS
+	SELECT s.*, cd.company_division, cr.company_regions
+	FROM staff s
+	LEFT JOIN company_divisions cd ON s.department = cd.department
+	LEFT JOIN company_regions cr ON s.region_id = cr.region_id;
+
+
+SELECT COUNT(*)
+FROM vw_staff_div_reg;
+
+
+/* How many staffs are in each company regions */
+SELECT company_regions, COUNT(*) AS total_employees
+FROM vw_staff_div_reg
+GROUP BY 1
+ORDER BY 1;
+
+
+SELECT company_regions, company_division, COUNT(*) AS total_employees
+FROM vw_staff_div_reg
+GROUP BY 1,2
+ORDER BY 1,2;
+
+
+/***** Grouping Sets *****************/
+-- Grouping Sets : allows to have more than one grouping in the results table
+-- there is no need to seperately use group by per query statement
+
+-- 2 groupings
+SELECT company_regions, company_division, COUNT(*) AS total_employees
+FROM vw_staff_div_reg
+GROUP BY 
+	GROUPING SETS(company_regions, company_division)
+ORDER BY 1,2;
+
+
+-- 3 groupings
+SELECT company_regions, company_division, gender, COUNT(*) AS total_employees
+FROM vw_staff_div_reg
+GROUP BY 
+	GROUPING SETS(company_regions, company_division, gender)
+ORDER BY 1, 2
+
+----------------------------------------------------------------------------------------------
+
+
+CREATE OR REPLACE VIEW vw_staff_div_reg_country AS
+	SELECT s.*, cd.company_division, cr.company_regions, cr.country
+	FROM staff s
+	LEFT JOIN company_divisions cd ON s.department = cd.department
+	LEFT JOIN company_regions cr ON s.region_id = cr.region_id;
+
+/* employees per regions and country */
+SELECT company_regions, country, COUNT(*) AS total_employees
+FROM vw_staff_div_reg_country
+GROUP BY 
+	company_regions, country
+ORDER BY country, company_regions;
+
+-------------- ROLL UP & CUBE to create Sub Totals ----------------------------------------
+-- both are sub clauses of Group by
+
+
+-------------- ROLL UP ----------------
+-- is used to generate sub totals and grand totals
+
+/* number of employees per regions & country, Then sub totals per Country, Then toal for whole table*/
+SELECT country,company_regions, COUNT(*) AS total_employees
+FROM vw_staff_div_reg_country
+GROUP BY
+	ROLLUP(country, company_regions)
+ORDER BY country, company_regions;
+
+
+----------- CUBE -------------------
+SELECT company_division, company_regions, COUNT(*) AS total_employees
+FROM vw_staff_div_reg_country
+GROUP BY 
+	CUBE(company_division, company_regions)
+ORDER BY company_division, company_regions;
+
+
+-------------------------------------------------------------------------------
+
+/* Difference between Roll Up and Cube 
+
+Reference: https://www.postgresqltutorial.com/postgresql-rollup/
+
+For example, the CUBE (c1,c2,c3) makes all eight possible grouping sets:
+(c1, c2, c3)
+(c1, c2)
+(c2, c3)
+(c1,c3)
+(c1)
+(c2)
+(c3)
+()
+
+
+However, the ROLLUP(c1,c2,c3) generates only four grouping sets, assuming the hierarchy c1 > c2 > c3 as follows:
+
+(c1, c2, c3)
+(c1, c2)
+(c1)
+()
+
+*/
+
+SELECT company_division, company_regions, country, COUNT(*) AS total_employees
+FROM vw_staff_div_reg_country
+GROUP BY 
+	ROLLUP(company_division, company_regions, country)
+ORDER BY company_division, company_regions, country;
+
+
+SELECT company_division, company_regions, country, COUNT(*) AS total_employees
+FROM vw_staff_div_reg_country
+GROUP BY 
+	CUBE(company_division, company_regions, country)
+ORDER BY company_division, company_regions, country;
+
+
+-------------------------------------------------------------------------------
+
+------------ FETCH FIRST ----------
+/* 
+Fetch First works different from Limit.
+
+For Fetch First, Order By Clause works first for sorting. Then Fetch First selets the rows.
+
+For Limit, Limit acutally limits the rows and then performs the operations.
+
+*/
+
+
+/* What are the top salary earners ? */
+SELECT last_name, salary
+FROM staff
+ORDER BY salary DESC
+FETCH FIRST	10 ROWS ONLY;
+	
+	
+/* Top 5 division with highest number of employees*/
+SELECT
+	company_division,
+	COUNT(*) AS total_employees
+FROM vw_staff_div_reg_country
+GROUP BY company_division
+ORDER BY company_division
+FETCH FIRST 5 ROWS ONLY;
+
+
+
+SELECT
+	company_division,
+	COUNT(*) AS total_employees
+FROM vw_staff_div_reg_country
+GROUP BY company_division
+ORDER BY company_division
+LIMIT 5;
+