Databases and DBMS: Models, Keys, SQL and Normalisation

The correct answer is A, hierarchical. The hierarchical model is the first data model used in commercial database software; IBM built its Information Management System on it in the 1960s for the Apollo programme. It arranges records as a tree in which every child record has exactly one parent, so data is reached by moving down a path from the root. Option B is wrong because the object oriented model, which stores objects together with their methods, is the newest of the four and belongs to the late 1980s. Option C is wrong because the deductive model, which adds logical rules to stored facts, came even later and remains mostly a research model. Option D is wrong because the relational model was proposed by Edgar Frank Codd in 1970, after the hierarchical and network models were already in use, though it later replaced both in practice.

The correct answer is B, Edgar Frank Codd. Codd, a British computer scientist working at IBM, set out the relational model in a paper published in 1970, proposing that all data be held in simple two-dimensional tables and joined by matching values rather than by fixed pointers. Every relational database, from Oracle to MySQL, rests on that idea, and he is therefore called the father of the relational database. Option A is wrong because Charles Babbage designed the Difference Engine and the Analytical Engine in the nineteenth century and is called the father of the computer. Option C is wrong because Dennis Ritchie developed the C language and, with Ken Thompson, the Unix operating system. Option D is wrong because James Gosling created the Java programming language at Sun Microsystems, which has nothing to do with the relational model.

The correct answer is C, tuple. In the vocabulary of the relational model a table is a relation, and one complete horizontal entry of that table, such as all the details of a single student, is a tuple; the ordinary words for it are row and record. Option A is wrong because an attribute is a column of the table, that is one property such as roll number or date of birth, and is counted by the degree of the relation. Option B is wrong because a domain is the set of values an attribute is allowed to take, for example the whole numbers one to one hundred for a marks column. Option D is wrong because the relation is the entire table, not one of its rows. Keep the pairing firm: relation is the table, tuple the row, attribute the column, and cardinality counts the tuples.

The correct answer is B, CREATE. The Data Definition Language deals with the structure of the database rather than with the data inside it, and CREATE is its basic command, used to build a table, a view or the database itself; ALTER, DROP, TRUNCATE and RENAME belong to the same group. Option A is wrong because INSERT adds new rows to an existing table and is therefore a Data Manipulation Language command. Option C is wrong because SELECT reads rows and returns a result set, which is again manipulation of data, not definition of structure; some books place it in a separate Data Query Language. Option D is wrong because UPDATE changes values already stored in rows and is also DML. Remember the split by what changes: structure means DDL, contents mean DML, rights mean DCL with GRANT and REVOKE.

The correct answer is C, TRUNCATE. TRUNCATE empties a table in one operation, leaving the columns, data types and constraints in place so that the table can be filled again at once; because it changes the object rather than selected rows it is grouped with the Data Definition Language and it is much faster than deleting row by row. Option A is wrong because DELETE is a Data Manipulation Language command that removes only the rows a WHERE clause selects and can be undone with ROLLBACK before the transaction is committed. Option B is wrong because DROP destroys the table itself, so the structure disappears along with the data. Option D is wrong because ERASE is not an SQL command at all; it is a distractor built from ordinary English. The DELETE against TRUNCATE difference is a favourite one-mark question.

The correct answer is C, primary key. The designer picks one candidate key as the primary key of a table, and the database then enforces two rules on it: its value must be different in every row, and it can never be left null, since a missing value would make a row impossible to identify. Option A is wrong because a foreign key holds the primary key values of another table to create a link, and it may be null when no related row exists. Option B is wrong because an alternate key is simply a candidate key that was not chosen as the primary key, so it is unique but is not the identifying key of the table. Option D is wrong because a super key is any set of attributes that identifies a row, and it may carry extra attributes that are not needed for uniqueness.

The correct answer is B, foreign key. A foreign key is the column that stores values belonging to the primary key of another table, so a marks table can carry the roll number of a student table and the two tables can then be joined; the database uses it to enforce referential integrity, refusing a value that does not exist in the parent table. Option A is wrong because a candidate key is a minimal set of attributes that identifies rows within the same table and points nowhere else. Option C is wrong because a composite key is a primary key built from two or more columns taken together, such as roll number with subject code. Option D is wrong because a secondary index only speeds up searching on a column and creates no relationship between tables at all.

The correct answer is C, isolation. Isolation means that transactions running at the same time do not see one another's incomplete work, so two clerks booking the last seat cannot both succeed; the database behaves as though the transactions ran one after the other. The four properties in full are atomicity, consistency, isolation and durability. Option A is wrong because integrity is a general term for correctness of data, enforced by constraints such as primary and foreign keys, and it is not the I of ACID. Option B is wrong because indexing is a technique for faster searching and has nothing to do with transaction guarantees. Option D is wrong because inheritance is a feature of object oriented programming, where a class takes properties from a parent class. Atomicity means all steps happen or none, and durability means a committed change survives a crash.

The correct answer is B, degree. The degree of a relation counts its attributes, that is its columns, so a student table holding roll number, name, class and marks has degree four whatever the number of students in it. Option A is wrong because cardinality counts the tuples, that is the rows of the table, so the same student table with two hundred students has cardinality two hundred. Option C is wrong because a domain is the set of permitted values for one attribute, for example the letters A to E for a grade column, and it is not a count at all. Option D is wrong because tuple count is only another name for cardinality, so it repeats the error in option A. Examiners frequently swap degree and cardinality in the options, so read the words attribute and tuple carefully before choosing.

The correct answer is B, second normal form. A table is in second normal form when it is already in first normal form and no non-key attribute depends on only a part of a composite primary key; such a partial dependency is removed by moving that attribute into a table of its own. Option A is wrong because first normal form only requires that every value be atomic and that repeating groups be removed, which says nothing about composite keys. Option C is wrong because third normal form removes transitive dependency, where one non-key attribute depends on another non-key attribute rather than on a part of the key. Option D is wrong because Boyce-Codd normal form is a stricter version of the third form, requiring every determinant to be a candidate key. Learn the order: 1NF atomic values, 2NF partial, 3NF transitive.

The correct answer is C, MongoDB. MongoDB stores data as documents in a flexible, JSON-like form instead of as rows in tables with a fixed set of columns, and it is queried through its own interface rather than mainly through SQL, which makes it the standard example of a NoSQL database in exam papers. Option A is wrong because Oracle Database is the best known commercial relational system and uses SQL. Option B is wrong because MySQL is an open source relational database, now owned by Oracle, and the M of the LAMP stack. Option D is wrong because Microsoft SQL Server is Microsoft's relational database and works through Transact-SQL. The clue in such questions is the word table: relational systems force a fixed table structure, while NoSQL systems do not.

The correct answer is C, DCL. The Data Control Language governs who may do what inside the database: GRANT gives a user a privilege such as the right to read or to update a table, and REVOKE withdraws a privilege already given. Option A is wrong because the Data Definition Language changes the structure of database objects through CREATE, ALTER, DROP and TRUNCATE. Option B is wrong because the Data Manipulation Language works on the rows themselves through INSERT, UPDATE, DELETE and SELECT. Option D is wrong because the Transaction Control Language manages a transaction as a unit with COMMIT to make changes permanent, ROLLBACK to undo them and SAVEPOINT to mark a place to roll back to. Fix the four groups by their purpose: structure, data, rights and transactions.