Assignment 1: Metadata Management

CS457/657 Database Management Systems
Programming Assignment 1: Metadata Management
Overview
In this assignment, you will write a program that allows a database user to manage the metadata of their relational
data. By metadata, we mean the database’s high-level information (e.g., database’s name, creation time, owner) as
well as the properties of the tables (e.g., table’s names, attributes, constraints).
System Design
• You are free (in fact, encouraged) to come up with your own design
o For instance, Sqlite3 uses one single file for each “database.”
• Here is one possible design:
o One Linux directory - a database
o One regular file - a table
Implementation
• The program should not use an external database library or an existing SQL parser/compiler.
• Any programming language is acceptable, e.g., Python, Java, C/C++, Go
o Please pick one that you are most comfortable/proficient with
o If you want to choose a language not mentioned above, please contact the TA before you
start coding
• Functionalities:
o Database creation, deletion
o Table creation, deletion, update, and query
Interface
• A similar but simpler interface than Sqlite3
• Examples (on a Linux terminal):
o # ./<your_program <enter
# CREATE DATABASE db_name <enter;
▪ The shell should prompt whether the command is successful or failed
• If failed, don’t crash but gracefully prompt why
▪ Then when you check your file system, it might look like this:
• ~/your_home/cs457/pa1/db_name
o # ./<your_program <enter
# USE db_1; CREATE TABLE test_tbl (a1 int, a2 char(9));
▪ If successful, then your file system might look like this:
• ~/your_home/cs457/pa1/db1/test_tbl
2 CS457/657 Database Management Systems, University of Nevada, Reno
Testing
• We will test your program on Ubuntu (version 14 or above)
• If your program cannot compile on our testbed, we may ask you to demo your program
o Hint: try not to use many exotic libraries
• A full test script will be provided
o # ~/cs457/pa1/<your_program < PA1_test.sql (expect the standard input)
▪ Alternatively, you can use a file name as an argument to your program.
▪ You will NOT lose points by only supporting a filename-argument interface, but
keep in mind that the standard input interface would be more desirable for your
users (e.g., our TA, Abdullah).
o # <expected output, hopefully…
o You don’t need to parse the comment lines (i.e., starting with “- -“)
o We will not to test your programs with any other scripts/commends
▪ However, it’s always good to consider more edge cases
▪ Try not to hardcode your parser:
• You want to parse them into a series of (dynamic) words
Grading (20 points total)
• This is an individual assignment.
• Design document that clarifies the followings: (5 points)
o How your program organizes multiple databases
o How your program manages multiple tables
o At a very high level, how you implement those required functionalities
• Source code (15 points)
o Coding style and clarity, 5 points
▪ Appropriate parenthesis locations, indention, etc.
▪ Always write comments at the beginning of any files
• Author, date, history, etc.
▪ Always write comments at the beginning of any non-trivial class/function
• What this class/function does, high-level algorithm if needed
▪ Write in-line comments for non-trivial blocks of code
o Functionality, 10 points
▪ Refer to the test script for detailed breakdowns
Submission
• WebCampus
• Compress all your source code and report into one package in this format:
o <your_netid_pa1
• Late penalty: 10% per day