Assignment 2: SQL

Due: 6:00pm Friday, 28 April 2023

Weighting: 10% of your total assessment

Research Papers Database

Description

You and a group of fellow undergrads have created a start-up like Google Scholar called ‘newScholar’ . newScholar is a free accessible web search engine that provides a broad range of information on scholarly  publications.  It  also  provides  information  on  researchers  and  relations  among scholarly publications and researchers.

For each researcher, newScholar records the researcher’s details such as first name, last name, and one email address. Each researcher can also be associated with a few keywords representing their ‘research area’, such as Databases, Machine learning, Psychology, Medicine, etc. For each researcher, newScholar maintains their list of publications. The researchers who author a publication together are called ‘co-authors’ .

For each publication, newScholar stores its title, date of publication, start page number (e.g., 475), end page number (e.g., 500), and a list of authors (there can be multiple authors of a publication, where each author is a researcher). Each publication can also be associated with a few keywords representing its ‘research area’, such as  Databases, Machine learning, Psychology, Medicine, etc.  NewScholar database will not store the actual publications, but rather a link to the document objects. Each publication is linked to one document object.  For each document object, newScholar stores the URL link and the document size in KB. Each publication has a list of references (i.e., it “cites” other publications), where each reference is another publication. If publication A is in the reference list of another publication B, then A is “cited by” B. Figure 1 shows an example publication with its basic information and its list of references.

NewScholar manually curates a list of top 10 publications every fortnight depending on the number of citations of the publications. A publication can make it to top 10 more than once over time. For each paper that is in the top 10 for a particular fortnight at a particular position, newScholar keeps a record of the start date when a publication reached that position (beginning of the fortnight) and end date (end of the fortnight). If a publication is again in the next fortnight’s top 10 (whether in the same position, or different), a new row is recorded, with the start/end dates being the start/end dates of this new fortnight. For example, the publication entitled “Learning to index” can be number 1 for the fortnight from January 1st - January 14th, 2022 but the rank drops to number 3 from January 15th - January 28th, 2022. This would result in 2x rows, one with position 1 and one with position 3. Note that we do not need to know the rationale why this is the current ranking.

Fig 1: An example of publication with title, authors, and the list of references

The Data Model

Fig 2: The physical ER model of newScholar startup database.

Questions:

1.   List  all  publications  with  no  references  (i.e.,  find  publications  that  do  not  cite  any  other publications). Your query should return results of the form (publicationID, title). (1 mark)

2.   Find the most recent publication. Assume there are no ties (only one publication is the most recent). Your query should return results of the form (publicationID, title, dateOfPublication). (1 mark)

3.   List all papers from the author “Renata Borovica-Gajic” that are at least 10 pages in length. Your query should return results of the form (publicationID, title). (1 mark)

4.   Find the publication with the highest number of citations. Note that, if a publication A is in the reference list of publication B and publication C, then the number of citations of A is 2. If there are ties, then you must return all publications with the highest number of citations. Your query should return results of the form (publicationID, title, citationCount), with one row per publication in case of a tie. (2 marks)

5.   List the URLs of the papers that have ‘Databases’ as one of their research areas and have been cited at least once. Your query should return results of the form (publicationID, documentUrl). (2 marks)

6.   Find the researchers who have at least 2 citations overall, and who have had publications in the top ten list at least once. Your query should return results of the form (firstName, lastName). (2 marks)

7.   Find which research area has had the highest number of publications in the top-ten publication list over time. If a publication has appeared in the top 10 several times, count such publications just once in the calculation. If there are ties, then you must return all results. Your query should return results of the form (keywordID, word), with one row per research area in case of a tie. (2 marks)

8.   List the researcher with the highest total citations. If there are ties, then you must return all such results. Your query should return results of the form (firstName, lastName, totalCitationCount), with one row per researcher in case of a tie. (3 marks)

9.   Find the researchers in ‘Databases’ research area who have co-authored at least one publication with at least one researcher from ‘Machine learning’ research area (regardless of the research area of the publication itself). Note that we only want to consider co-authorships where the ‘Databases’  researcher  is  distinct  from  the  ‘Machine  Learning’  researcher:  if  there  is  a ‘Databases’ researcher who also is in the ‘Machine learning’ area, they must have co-authored with a different researcher who is in the ‘Machine learning’ area to be included. Your query should return results of the form (firstName, lastName). (3 marks)

10. Find the researchers who have not co-authored a publication on or after 01/01/2023 (where 01/01/2023 is the date of the publication) with the researcher named “Renata Borovica-Gajic” but have had at least one publication before that date co-authored with “Renata Borovica-Gajic” (that means they have had joint publications in the past, but not on or after 01/01/2023). Your

query should return results of the form (firstName, lastName) for all such researchers. (3 marks)

SQL Response Formatting Requirements

To help us mark your assignment queries as quickly/accurately as possible, please ensure that:

•   Your query returns the projected attributes in the same order as given in the question, and   does not include additional columns. E.g., if the question asks ‘return as (userId, name)’ ,     please write “SELECT userId, name …” instead of “SELECT name, userId…”  (you can name the columns using `AS` however you’d like, only the order matters).

•    Please do NOT use “databaseName.tableName” format. E.g., please write “SELECT userId FROM users…” instead of “SELECT userId FROM coltonc.users …” .

•    Ensure that you are using single quotes( ‘ ) for strings (e.g.  …WHERE name = ‘bob’…)and double quotes ( “ ) only for table names (e.g. SELECT name FROM “some table name with spaces”… ). Do NOT use double quotes for strings “…WHERE name = “bob”…” .

Submission Instructions

Your submission will be in the form of an SQL script. There is a template file on the LMS, into which you will paste your solutions and fill in your student details (more information below).

This .sql file should be submitted on Canvas by 6pm on the due date of Friday, 28 April 2023. Name your submission as 987654.sql, where 987654 corresponds to YOUR student id.

Filling in the template file:

The template file on the LMS has spaces for you to fill in your student details and your answers to the questions. There is also an example prefilled script available on the LMS as well. Below are screenshots from those two documents explaining the steps you need to take to submit your solutions:

Late submission

Unless you have an approved extension (see below), you will be penalised - 10% of the total number of marks in the assignment per day that your submission is late. For instance, if you received a 78% raw score, but submitted 2 days late, you'd receive a 58% score for the assignment.