Code-Reading and Design Questions

General guidelines

These questions are designed to help you to prepare for your work on the problems from Part II of Problem Set 3.

Before you begin

You should have already performed the tasks in the Getting started section of the problem set. In particular, you should have read the overview of the code framework.

While answering the questions below, you should consult the following resources:

• the Java files we have provided in the dbms code framework
• the overview of the code framework
• the API documentation of the code framework
• the lecture notes on implementing a logical-to-physical mapping
• the Berkeley DB Java Edition API

The Table and Column classes

The Table class is a blueprint for objects that represent one of the tables in the database. The Column class is a blueprint for objects that represent one of the columns in a table.

Review:

• the API of the Table and Column classes
• the actual code in those classes.

and answer the following questions:

1. What does a Table object’s open() method do? If you try to call this method for a table that doesn’t exist, what does the method return? How should your code deal with that return value?

2. What Table method allows you to determine whether a table has a primary key?

3. What Table method allows you to obtain the columns associated with the table? What method call would you use to obtain the leftmost column in the table?

4. What Column method allows you to determine the data type of the column? What are its possible return values?

5. What Column method allows you to determine the length of the values that will go in that column? What does it return for VARCHAR columns?

The SQLStatement class and its subclasses

SQLStatement is an abstract superclass that includes the fields and methods needed by one or more types of SQL commands. Each type of SQL command is represented by a different subclass of SQLStatement, and the subclasses inherit much of their functionality from SQLStatement.

Review:

• the API of the SQLStatement class
• the actual code in that class

and answer the following questions:

6. What method allows you to determine the table or tables on which a given SQL command should operate?

7. What methods allow you to obtain information about the other components of the command (e.g., the columns)?

Marshalling

As part of this assignment, you will implement code that marshalls one row of a table into a key/value pair. You will use the approach covered in lecture in which:

• The key portion of the key/value pair holds the value of the primary-key column.

• The value portion of the key/value pair begins with a table of offsets that is followed by the values of all columns except the primary key.

You will add code to the provided InsertRow class to perform the necessary marshalling. Every InsertRow object includes fields for two RowOutput objects, which allow you to write values into an underlying buffer (i.e., a byte array).

Review:

• the lecture notes on variable-length records that begin with a header of offsets
• the lecture notes on marshalling
• the API of the InsertRow and RowOutput classes
• the actual code in those classes

and answer the following questions:

8. Every column will have an offset in the offset table. This includes the primary key–even though it will be stored in the value portion of the key/value pair. What offset should you use for the primary key?

9. What offset should you use for a NULL value?

10. Consider the following sequence of SQL commands:

    CREATE TABLE Foo (a INT PRIMARY KEY, b VARCHAR(20));
    INSERT INTO Foo VALUES (1, 'hello');
    INSERT INTO Foo VALUES (2, NULL);
    

    Describe the key/value pairs that the marshalling scheme would create for the two rows inserted by the INSERT commands above. For each row, what will the key look like and what will the value would look like?

11. Now consider the following sequence of SQL commands:

    CREATE TABLE Bar (a INT, b CHAR(4) PRIMARY KEY, c VARCHAR(15), d REAL);
    INSERT INTO Bar VALUES (1, '1234', 'hello', 12.5);
    INSERT INTO Bar VALUES (2, '4567', 'wonderful', NULL);
    

    Describe the key/value pairs that your marshalling scheme would create for the two rows inserted by these INSERT commands.

12. The RowOutput methods that you will use for writing offsets and column values are inherited from the DataOutputStream class, so you should make sure to review the API of that class.

    Which methods of this class will you use, and for what purpose will you use each method?

Unmarshalling

You will add code to the provided TableIterator class to perform the necessary unmarshalling. To do so, you will make use of RowInput objects, which allow you to read values from an underlying buffer (i.e., a byte array).

Review:

• the lecture notes on unmarshalling
• the API of the TableIterator and RowInput classes
• the actual code in those classes

and answer the following questions:

13. How will you retrieve a particular column value from a row? In other words, will you go from a key/value pair to the value of a specific column?

14. What methods from the RowInput class will you use to read a value from the buffer that contains the marshalled values?

15. When unmarshalling, how will you determine the length of a VARCHAR attribute so that you can read in the appropriate number of bytes?

Miscellaneous

16. In InsertStatement.java, we have given you the beginning of the execute() method for INSERT statements. How does the code that we have given you make use of objects from the InsertRow class and their associated methods?

17. What remains to be done in the rest of the execute() method, and what Berkeley DB methods will you need to use in the code that you write to complete that method?

18. In executing a SELECT * statement, you will make use of the printAll() method for table iterators. We have given you this method in the TableIterator class, but it invokes a method that you will need to implement. What method is it?

Last updated on October 19, 2024.

