Introduction

This is an individual assignment. You will complete 4 parts with 7 questions (total marks 25) to demonstrate  your  understanding  of  multiple  topics,  including  distributed  database,  data warehousing, data integration, and data quality management. Meanwhile, coding is required for some questions to show your problem-solving ability.

Notes:

1.   It is highly suggested to complete Prac 3 before working on the coding part of this assignment (i.e., for Part 4 of this assignment). Although the assignment is independent from the three practicals, the code introduced in Prac 3 can be the starting point of this assignment as the tasks are similar.

2.   Each dataset used in this assignment contains thousands of records, which is hard to be checked record-by-record manually. Therefore, it is recommended to have a handy text editor tool (e.g. Microsoft Excel, Notepad++ or Sublime Text on Windows) to view and search the contents in CSV files. Please use search function (i.e.,  Ctrl+F keys) in text editor to look through values. Also, please don’t change the data unintentionally while viewing or searching, as it may affect your assignment results.

3.   Implement your code in SQL, Java or Python, choose the one you feel comfortable. The code must be companied by minimum comments so that tutors can understand the structure of your coding and the objective of each snippet.

Assessment Requirements:

Please include all your answers in a word/pdf document. Pack the documents with your code folder (which contains at least “src” and “data” folders, shown as below) into a .zip/.rar file and submit it to the Blackboard INFS3200 course Website. The name of both the zip file and the document should contain your student ID, your name and “Assignment”, shown as follows:

Please format your document nicely, in terms of consistent font, font size and spacing. The answers are suggested to follow the below structure (No need to repeat questions ifnot necessary, fonts and spacing are not limited):

…

Part 1.

Question 1: Your answers …

Question 2: Your answers …

Part 2.

…

WARNING: This assignment must be completed individually. Any form of answer-sharing with other people is not acceptable and, once identified, will be penalized.

Preliminary: Dataset Description

In this assignment, we have four datasets about book information from four different sources. The data schemas are listed below:

Book1 (id, title, authors, pubyear, pubmonth, pubday, edition, publisher, isbn13, language, series, pages)

Book2 (id, book_title, authors, publication_year, publication_month, publication_day, edition, publisher_name, isbn13, language, series, pages)

Book3    (ID,    Title,   Author1,   Author2,   Author3,   Publisher,   ISBN13,   Date,   Pages, ProductDimensions, SalesRank, RatingsCount, RatingValue, PaperbackPrice, HardcoverPrice, EbookPrice, AudiobookPrice)

Book4 (ID, Title, UsedPrice, NewPrice, Author, ISBN10, ISBN13, Publisher, Publication_Date, Pages, Dimensions)

Part 1: [6 marks] Database Schema and Fragmentation

Read the above schemas carefully and understand the meaning of the attributes. If you don’t know the meaning of a certain attribute, check the data under it or Google its meaning (especially for   some   abbreviations,  like  ISBN).  Answer  the   following  questions  based  on  your understanding.

Question 1: [2 marks] Given four datasets that are stored in one relational database as separate relations.

(1)    Write an SQL query “Find the top 5books that have the  highest ratings and 5 books that have the lowest ratings, return their ranks (sorted in descending order), titles, publishers and number of pages” .

(2)    Which table schema(s) is/are used to answer the above query?

Question 2: [4 marks] Given that Book3 is stored in a distributed database A, and two queries that are most frequently asked on A are:

  Find all books whose publisher name is “XXX” (or among multiple publishers), return

their book titles and author info.

  Find  all  books  that  are  published  in  a  given  year,  return  their  book  IDs,

languages,number of pages, HardcoverPrice and EbookPrice.

Answer the following questions:

(1)     [2 marks] If the goal ofA is to handle each query by a dedicated local site (no information needed from the other site), which fragmentation strategy should be used to fragment Book3 table? If only two fragments are generated, write their schemas  (if vertically fragmented) or predicates (if horizontally fragmented), respectively. (Note: there are lots of valid fragmentation solutions, just provide one of them.)

(2)      [2 marks] Assuming that we horizontally fragment the table into three fragments based on the following predicate:

Fragment 1: pages ≤ 250

Fragment 2: 250 < pages ≤ 750

Fragment 3: pages > 1000

Is this set of predicates valid? If so, please explain (using plain English) the insertion process if we want to insert a new record into Book3. If not, please generate a valid predicate set using minterm predicates (show the calculation process). Also, explain the insertion process for a new record after the valid predicate set is made.

Part 2: [7 marks] Data Warehouse Design

In this part, we design a Data Warehouse on book sales w.r.t. the Book1, Book2, Book3, and Book4 datasets. Particularly, we need to use data from the given assignment datasets and create a Data Warehouse Schema. The designed Data Warehouse will contain summary data, such as the total sales of each publisher, for each day and each language. The following shows just an example:

Question 3: Design a Data Warehouse Schema that can accommodate the above example, answer the following questions:

(1)     [1 mark] Show the schema and point out the dimensions and fact table. Given that we have a dimension table for each dimension and there are 4000 records in the fact table. Among all dimension tables and the fact table, which table has the most records? Why?

Question 4: Now we want to the create bitmap indices for the given model:

(1)     [2 marks] What are the advantages of building a bitmap index? Which type of column is not suitable for bitmap index?

(2)     [2  marks]  Suppose  the  “Publisher”  column  only  contains  four  distinct  values  and “Language” only contains two, which are all shown in the above example. Please create bitmap indices for both “Publisher” and “Language” .

(3)     [2 marks] Explain how to use the bitmap indices to find the total sales of “English” books published by “AAAI Press” .

Part 3: [4 marks] Data Integration

Given that the data warehouse loads data from the above four sources (Book 1,2,3,4), you are asked to integrate their data and address various data quality issues. In this part, those database sources (i.e., owners) only give you their schemas (shown in Preliminary part), and you are asked to design an integrated schema based on the given schemas (i.e., the data records within tables Book 1,2,3,4 are supposedly not available for you at this stages).

Question 5: Now you define a global schema (using the approach namely, Global as a View) which can integrate data from all four sources.

(1)     [2 marks] Design a global schema which will combine the common attributes from each schema together. Your design should include any information that is represented in all four schemas. If an attribute cannot be found or derived in the given schemas, then it should be left out of your global schema.

(2)     [1 marks] Identify  structural heterogeneity issues that may occur during your integration by an example in the schemas together with the possible resolution.

(3)     [1 marks] Identify  semantic heterogeneity issues that may occur during your integration by an example in the schemas together with the possible resolution.

Part 4: [8 marks] Data Quality Issues

Now assume you are provided with the actual data from each source, namely “Book1.csv”, “Book2.csv”, “Book3.csv” and “Book4.csv” (see the Assignment provided datasets). As it is very common that the same book is recorded by different sources, it is crucial to identify the redundant  information  by  merging  and  eliminate  the  duplicated  records  during  the  data integration process, which relies on the data linkage techniques to be used. In this regard, we provide a human-labelled gold-standard dataset (refer to Prac 3 Part 2.2 for more information about  gold-standard),  named  as  “Book1and2_pair.csv”,  which  lists  all  correct  matchings between Book1 and Book2. It will be used in the following tasks. Its schema is as follows:

Book1and2_pair (Book1_ID, Book2_ID)

In a CSV file, you need to note that the attributes are separated by comma (,). If two commas appear consecutively, it means the value in the corresponding field between two commas is NULL (i.e., absent). Furthermore, if an attribute field contains comma naturally, the field will be enclosed by a double quote ("") to differentiate the actual comma notation inside attribute from the outside comma separator. For example, a record in Book2 is as follows:

1725,Informix  Unleashed,"John  McNally,  Jose  Fortuny,  Jim  Prajesh,  Glenn  Miller",

97,6,28,1,Sams,9.78E+12,,Unleashed Series,1195

According to Book 2 schema, we can infer the following fields:

id=1725,

book_title=Informix Unleashed,

authors= John McNally, Jose Fortuny, Jim Prajesh, Glenn Miller,

…

isbn13=9.78E+12

language=NULL,

series=Unleashed Series,

pages=1195.

Here, since there are commas in “authors” field, the whole field is enclosed by a notation of double quotes. Also, since there are two consecutive commas before “Unleashed Series”, it means that the language is NULL.

In this part, you are asked to answer the following questions by writing code to complete the tasks (if “code required” is specified) and provide your answers based on the code results. Please store all the code you wrote during this part and submit them to Blackboard Course Website as a part of your assignment submission.

Question 6: Sample records from “Book3.csv” to measure its data quality:

(1)     [1 mark] By sampling the records whose id is the multiple of 100 (i.e. 0, 100, 200, 300, …), how many records are there in the sample set (code required)?

(2)     [1 mark] Among the samples found in Question 6-(1), how many fields containing NULL values are presented (code required)?