Question

Relational algebra (RA)

For the database schema shown below, write RA query for each of the following. (State any reasonable assumptions only if necessary)

employee (person_name, street, city)

works(person_name, company_name, salary)

company (company_name, city)

Questions:

1. Find person names, streets and cities of employees who work for two different companies.
2. For each company, find person name, salary, street and city for the highest paid employee of that company.
3. For each company, find total number of employees and total number of cities their employees reside.
4. Find name and city of the company which has highest-paid employee among employees from all companies, along with the name and salary of this employee.
5. Find name and city of any company which has employees living in Lincoln and Omaha. [Meaning of the query: For each of these companies, it has some employees live in Lincoln, and some others live in Omaha. Note no employee can live in to different cities.]

Answer 1

Answer:--------------

a.

b. cmm --> Πcompany_nameG max(salary) as s (works)
Result --> ( Π_{company_name, person_name, salary}(cmm |X| works |X| employee))

c. (Π_{Company_name}G _{count(person_name), count(distinct city})(works |X| employee)

d. (Π_{person_name, salary}((G _{max(salary) as maxS}(works)) |X| works))|X| Company

e. The easiest way is to use divide and conquer: Write two subquries: “Find name and city of any company which has employees living in Lincoln” and “Find name and city of any company which has employees living in Omaha,” then take intersection.Note: If you write σ_{employee.city = ‘Lincoln’ employee.city = ‘Lincoln’…} It will not work, because an employee can only live in one city, so the result is ALWAYS empty.
On the other hand, if you use the following to get company names, it does not guarantee correct answer, since companies with employees living in one of the two cities will also be returned:
σ_{employee.city = ‘Lincoln’ employee.city = ‘Lincoln’…}

To remedy this problem, some students included post-processing steps: first project out employee city names for each company (which should return two tuples including names of both cities), then take the division. Another approach is to count the number of employee cities for each company, and return the number of company name if the count is equal to two. Both approach are OK but less straightforward.