联系方式

  • QQ:99515681
  • 邮箱:99515681@qq.com
  • 工作时间:8:00-23:00
  • 微信:codinghelp

您当前位置:首页 >> Java编程Java编程

日期:2020-11-15 08:32

CSC173: Project 4

The Relational Data Model

In this project you will gain experience with databases and relational algebra by implementing

your own database system. We will build on the presentation (and code) in the

textbook.

Note: Your writeup for this project will probably be more extensive than previous projects.

In addition to the usual instructions for building and running your project, please be sure

to describe what your program is doing and what we’re seeing. Of course, that should

also be clear from the output of the program by itself.

Part 1

1. Implement a database containing the relations (tables) shown in FOCS Figure 8.1

and 8.2 (also seen in class). Use the method described in Section 8.2 in the

section “Representing Relations” and elaborated in Section 8.4 “Primary Storage

Structures for Relations.” Use hashtables. Describe your implementation briefly

but clearly in your writeup.

2. Implement the basic single-relation insert, delete, and lookup operations as functions.

If you use the implementation described in the textbook, you will need separate

functions for each relation (e.g., insert CSG). You should support leaving

some attributes unspecified for delete and lookup (denoted with “*” in the textbook,

perhaps something else in your code). Describe your implementation briefly but

clearly in your writeup.

3. Use your insert method to populate the tables with (at least) the data given in the

figures. Then demonstrate all three operations by performing the operations shown

in Example 8.2 (p. 409). Be sure that your program explains itself when run (using

informative printed messages).

1

Part 2

For this part, use the database with all the tuples from Figures 8.1 and 8.2 (that is, before

any deletions).

1. Write a function to answer the query “What grade did StudentName get in CourseName?”

as described in Section 8.6 “Navigation Among Relations.” The code for

your function MUST look like the pseudocode in Fig. 8.8. (I recommend using the

pseudocode as comments in your code.) Demonstrate this functionality by asking

the user for the query parameters, performing the query, and printing the results

informatively (that is, a kind of REPL).

2. Write a function to answer the query “Where is StudentName at Time on Day?”

(assuming they are in some course). Demonstrate this functionality also.

Part 3

1. Implement the Relational Algebra operations as described in Section 8.8 and demonstrate

this by doing the operations on the “registrar” database described in Examples

8.12 (Selection), 8.13 (Projection), 8.14 (Join), and 8.15 (all three). Describe

your implementation briefly but clearly in your writeup.

? If you follow the implementation in the textbook, you will probably need a

different function for each different set of arguments to the operator (e.g.,

select CSG, join CSG SNAP). Only implement the ones that you need for

the examples listed above. You may also throw in any additional examples

that you feel illustrate relevant aspects of your implementation.

? Note that some of the Relational Algebra operations create a relation with a

new schema from that of their operands. (You should know which operations

do this. . . ) But if tuples are implemented using structs, how do you create

instances of these new “on-the-fly” relations?

The answer is that you should solve the problem yourself by hand so that

you know the schemas needed by your examples. Then add appropriate

structure definitions to your program to allow you to do the examples. This is

one of many differences between what you need to do for this project and a

real database framework.

2

Extra Credit (20% max total)

1. Implement functions for saving your database to one or more files, and loading

from the same. Demonstrate this functionality in your program(s) and explain in

your writeup. Note: you can do all the other parts of the project even without this

functionality, so don’t let it stop you or slow you down. [max 10% extra]

2. The code you wrote for the “registrar” database is obviously specific to it. A true

database system, like SQLite or MySQL, allows you to represent any database

schema. Generalize your code to represent arbitrary databases consisting of arbitrary

relations. Note that you do not need to understand SQL for this. What you

need to do is create “generic” representations of tuples and tables and then use

them to write code for some specific example.

If you choose to do this, and are confident in your implementation, you may use it

for the required parts of the project. Explain what you’ve done in your writeup and

we will give you the extra points if it all works. You may also choose to demonstrate

this separately, for example with another program. Again, be clear in your writeup

so that we can give you the points. [max 10% extra]

3. Identify a subset of SQL that you can support with your implementation. You should

focus on the query parts of the language, not the data definition parts. Write a

simple parser for that language, and use it to query (and perhaps also create) the

“registrar” database. Demonstrate this by also allowing the user to enter queries

and have your system parse and execute them and print the results. Explain what

you’ve done and how it will be demonstrated by the program(s) in your writeup.

[max 10% extra]

Project Submission

Your project submission MUST include the following:

1. A README.txt file or README.pdf document describing:

(a) Any collaborators (see below)

(b) How to build your project

(c) How to run your project’s program(s) to demonstrate that it/they meet the

requirements

2. All source code for your project. Eclipse projects must include the project settings

from the project folder (.project, .cproject, and .settings). Non-Eclipse

3

projects must include a Makefile or shell script that will build the program per

your instructions, or at least have those instructions in your README.txt.

3. A completed copy of the submission form posted with the project description.

Projects without this will receive a grade of 0. If you cannot complete and save

a PDF form, submit a text file containing the questions and your (brief) answers.

We must be able to cut-and-paste from your documentation in order to build and run

your code. The easier you make this for us, the better grade your will be. It is your

job to make both the building and the running of programs easy and informative for your

users.

Programming Policies

You must write your programs using the “C99” dialect of C. This means using the “-std=c99”

option with gcc or clang. For more information, see Wikipedia.

You must also use the options “-Wall -Werror”. These cause the compiler to report

all warnings, and to make any warnings into errors that prevent your program from compiling.

You should be able to write code without warnings in this course.

With these settings, your program should compile and run consistently on any platform.

We will deal with any platform-specific discrepancies as they arise.

If you submit an Eclipse project, it must have these settings associated with the project.

Projects with that compile with warnings will be considered incomplete.

Furthermore, your program should pass valgrind with no error messages. If you don’t

know what this means or why it is A Good Thing, look at the C for Java Programmers

document which has a short section about it. Programs that do not receive a clean report

from valgrind have problems that should be fixed whether or not they run properly. If

you are developing on Windows, you will need to look for alternative memory-checking

tools.

Late Policy

Late projects will not be accepted. Submit what you have by the deadline. If there are

extenuating circumstances, submit what you have before the deadline and then explain

4

yourself via email.

If you have a medical excuse (see the course syllabus), submit what you have and

explain yourself as soon as you are able.

Collaboration Policy

You will learn the most if you do the projects YOURSELF.

That said, collaboration on projects is permitted, subject to the following requirements:

? Groups of no more than 3 students, all currently taking CSC173.

? You must be able to explain anything you or your group submit, IN PERSON AT

ANY TIME, at the instructor’s or TA’s discretion.

? One member of the group should submit on the group’s behalf and the grade will

be shared with other members of the group. Other group members should submit

a short comment naming the other collaborators.

? All members of a collaborative group will get the same grade on the project.

Academic Honesty

Do not copy code from other students or from the Internet.

Avoid Github and StackOverflow completely for the duration of this course.

There is code out there for all these projects. You know it. We know it.

Posting homework and project solutions to public repositories on sites like GitHub is a violation

of the University’s Academic Honesty Policy, Section V.B.2 “Giving Unauthorized

Aid.” Honestly, no prospective employer wants to see your coursework. Make a great

project outside of class and share that instead to show off your chops.

5

Frequently Asked Questions

Q: In C, how can I make a hashtable that can hold any kind of tuple?

A: How would you do it in Java? Think about it. . .

Answer: Generics. And indeed java.util.Hashtable is a generic class.

Now, does C have generics? Answer: no. So this is an example of the

kind of thing added to Java as it evolved from C, in order to capture common

programming patterns more effectively. But if all we have is C, what can we

do?

For some dynamic data structures, you can just manage “generic” objects of

type void* and use casts as needed (I STRONGLY suggest that you encapsulate

the casts in helper functions).

The problem is that unlike a linked list, a hashtable needs to know something

about the objects it is managing. Why? Because it has to hash them,

which means accessing some of their components. So you need to be able

to tell the “generic” hashtable code which hash function to use for each type

of struct it will be managing. There are ways (e.g., function pointer in tuple or

hashtable), and they all amount to reinventing (in fact, pre-inventing) the idea

of true classes.

You could just cut and paste the code for different flavors of hashtable. So the

routines use the right type of struct for arguments and they know what hash

function to use. Is this ugly? Yes. Is it bad software engineering practice?

Yes. Is it ok for this project? Yes. This is similar to the business about the

schemas of tuples produced by JOIN operations, as described in the project

description Part 3.

If you were ambitious you might observe that the cut-and-paste method is

purely mechanical, other than writing the hash function itself. That is, you

basically go through and change all occurrences of the element type name

and that’s really all you need to do. Given that, you could define a macro or

macros that expanded into the code. Something like:

#define RELATION(ETYPE,HFUNC) ...

for element type ETYPE and name of hash function HFUNC. Using macros

is a bit of a power C programmer technique. But it is how we did “generic”

code back in the day, and as you may know, C++ grew out of just this kind of

macro-powered code.

You could also implement tuples using something other than structs, as described

in one of the Extra Credit problems. There are many possibilities. . .

6


版权所有:编程辅导网 2018 All Rights Reserved 联系方式:QQ:99515681 电子信箱:99515681@qq.com
免责声明:本站部分内容从网络整理而来,只供参考!如有版权问题可联系本站删除。