The Relational Model What DB models exist? Most widely current model: Relational Model Vendors: IBM DB2, Microsoft, Oracle, Sybase, etc… “Legacy systems” have older models E.g., IBM’s IMS (hierarchical), CODASYL network model Recent and future competitors: object-oriented model: ObjectStore, Versant Object-relational model: Oracle, DB2 Semi-structured: XML Integrated in all major relational database systems Native XML database systems 421B: Database Systems - The Relational Model 2 1 Definitions Relational Database: a set of relations Relation: Consists of two parts: Schema: specifies name of relation, plus a set of attributes, plus the domain/type of each attribute E.g., Students(sid:string, name:string, login:string, faculty:string, gpa:real) Instance: set of tuples (all tuples are distinct). A relation can be seen as a table: Column headers = attribute names, rows = tuples/records, columns/fields = attribute values #Rows = cardinality #Fields = degree / arity If clear from context, we say instead of “instance of a relation” simply “relation” Database Schema: collection of relation schemas 421B: Database Systems - The Relational Model 3 Example of Students Relation sid name login faculty gpa Column headers = attributes 53666 Bartoli bartoli@cs Science 3.4 53688 Chang chang@eecs Eng 3.2 Cardinality = 3 53650 Chang chang@math Science 3.8 Degree = 5 ... All rows are distinct (set-oriented) Rows are not ordered (a permutation of rows represents still the same table) Columns are per definition not ordered but in practice we often assume a fixed order with this, a single tuple can be represented as (53666, Bartoli, bartoli@cs, Science, 3.4) 421B: Database Systems - The Relational Model 4 2 Relational DDL and DML Data Definition Language (DDL): defines the schema of a database Data Manipulation Language (DML): “manipulates” the data, i.e., the instances of the relations Insert, update, delete tuples “Query” the relations: retrieve tuples that fulfill certain criteria (hence, often called “query language”) The Relational Model offers simple and powerful querying of data with precise semantics independent of how data is stored or whether changes in the physical structure are made (physical data independence) 421B: Database Systems - The Relational Model 5 The SQL Query Language Developed by IBM (system R) in the 1970s Need for a standard since it is used by many vendors Standards: SQL-86 … SQL-99 / SQL3 (adds object-relational features) SQL:2003 (adds XML features) 421B: Database Systems - The Relational Model 6 3 SQL Data Types All attributes must have a data type. SQL supports several basic data types Character and string types CHAR(n) denotes a character string of fixed length (containing trailing blanks for padding if necessary). VARCHAR2(n) denotes a string of up to n characters (between 0 and n characters). SQL permits reasonable coercion between values of character- string types Integer Types INT or INTEGER (names are synonyms) SHORTINT 421B: Database Systems - The Relational Model 7 Data Types (contd.) Floating point numbers FLOAT or REAL (names are synonyms) DOUBLE PRECISION DECIMAL(n,d): real number with fixed decimal point. Value consists of n digits, with the decimal point d positions from the right. Dates and time: DATE: has the form ‘YYYY-MM-DD’ TIME: has the form ‘15:00:02’ or ‘15:00:02.5’ May be compared and converted to string types Bit strings User defined domains New name for a data type Possibility to define restrictions on values of domain (< 10) 421B: Database Systems - The Relational Model 8 4 Data Definition: Table Creation Defines all attributes of the CREATE TABLE Students relation (sid CHAR(20), The type/domain of each name VARCHAR2(20), attribute is specified DBMS enforce correct type login CHAR(10), whenever a tuple is added or faculty VARCHAR(20), modified gpa REAL SQL is case insensitive DEFAULT 0.0) It is possible to define default values CREATE TABLE Enrolled Special NULL value: ‘unknown’ (sid CHAR(20), cid CHAR(20), grade CHAR(2)) 421B: Database Systems - The Relational Model 9 Data Definition: Destroying and Altering Relations DROP TABLE Students Destroys the relation Students. The schema information and the tuples are deleted. ALTER TABLE Students ADD COLUMN firstYear:integer The schema of students is altered by adding a new field; every existing tuple in the current instance is extended with a null value in the new field. 421B: Database Systems - The Relational Model 10 5 Data Manipulation: insert, delete, update Insert a single tuple using: INSERT INTO Students (sid,name,login,faculty,gpa) VALUES(53688,’Chang’,’cheng@eecs’,’Eng’,3.2) INSERT INTO Students (sid,name,login,faculty) VALUES(53688,’Chang’,NULL,’Eng’) Can delete all tuples satisfying some condition DELETE FROM Students WHERE name = ‘Chang’ Can update all tuples satisfying some condition UPDATE Students SET gpa = 3.4 WHERE sid = 53688 421B: Database Systems - The Relational Model 11 Querying the Data Find the names and gpa of all students with gpa less than 3.5 SELECT name, gpa FROM Students WHERE gpa < 3.5 sid name login faculty gpa name gpa 53666 Bartoli bartoli@cs Science 3.4 53688 Chang chang@eecs Eng 3.2 Bartoli 3.4 53650 Chang chang@math Science 3.8 Chang 3.2 421B: Database Systems - The Relational Model 12 6 Integrity Constraints (ICs) Integrity Constraints must be true for any instance of the database; e.g., domain constraints ICs are specified when schema is defined ICs are checked when relations are modified A legal instance of a relation is one that satisfies all specified ICs. DBMS should not allow illegal instances If DBMS checks ICs, stored data is more faithful to real-world meaning (also checks for entry errors) Of course, DBMS can only check what is specified in the schema 421B: Database Systems - The Relational Model 13 Primary Key Constraints A set of fields is a key for a relation if No two distinct tuples can have same values in all key fields, and This is not true for any subset of the key. If there are two or more keys, one of the candidates is chosen to be the primary key. The primary key attributes of a tuple may not be NULL. A set of fields that contains a subset of fields fulfilling the key constraint is called a superkey E.g. sid is a key for Students. (What about name?). The set (sid,gpa) is a superkey 421B: Database Systems - The Relational Model 14 7 Primary and Candidate Keys in SQL Possibly many candidate keys exist, one of which is chosen as CREATE TABLE Students the primary key (sid CHAR(20) PRIMARY KEY, Each student has a unique id. name VARCHAR2(20), For a given student and course, … there is a single grade; further, no two students in a course receive the same grade CREATE TABLE Enrolled Application dependent (sid CHAR(20), Defined carelessly, an IC can cid CHAR(20), prevent the storage of database grade CHAR(2), instances that arise in practice! PRIMARY KEY (sid,cid), UNIQUE (cid,grade)) 421B: Database Systems - The Relational Model 15 Foreign Key Foreign Key: Set of fields in one relation that is used to “refer” to a tuple in another relation. Must correspond to the primary key of the second relation. Represents a “logical pointer”. Examples in relation Enrolled, sid is a foreign key referring to Students: Students(sid:CHAR(20),name:VARCHAR(20),login:CHAR(10), faculty:VARCHAR(20), gpa:REAL) Enrolled(sid:CHAR(20),cid:CAHR(20),grade:CHAR(2)) 421B: Database Systems - The Relational Model 16 8 Referential Integrity Foreign Key Constraint: the foreign key value of a tuple must represent an existing tuple in the referred relation Enrollment may only contain a tuple of a student who exists in the Students relation If all foreign key constraints are enforced, referential integrity is achieved, i.e., no dangling references 421B: Database Systems - The Relational Model 17 Foreign Keys in SQL CREATE TABLE Enrolled Only students listed in (sid CHAR(20), the Students relation cid CHAR(20), should be allowed to grade CHAR(2), enroll for courses PRIMARY KEY (sid,cid) FOREIGN KEY (sid) REFERENCES Students) sid cid grade sid name login faculty gpa 53666 Topology112 C 53666 Reggae203 B 53666 Bartoli bartoli@cs Science 3.4 53688 Chang chang@eecs Eng 3.2 53650 Topology112 A 53650 Chang chang@math Science 3.8 53668 History105 B 421B: Database Systems - The Relational Model 18 9 Enforcing Referential Integrity An Enrolled tuple with a sid is inserted such that no tuple with this sid exists in Students Disallow insertion A Students tuple is deleted Delete all Enrolled tuples that refer to it Disallow the deletion of a Students tuple to which Enrolled tuples point Set sid in Enrolled tuples that refer to it to “default sid” (in SQL set sid in Enrolled tuples that refer to it to NULL value) The primary key of a Students tuple is changed Update the sid of all Enrolled tuples that refer to the original value Further options similar to delete 421B: Database Systems - The Relational Model 19 Referential Integrity in SQL/92 SQL standard supports all CREATE TABLE Enrolled 4 options on deletes and (sid CHAR(20), updates cid CHAR(20), Default is NO ACTION grade CHAR(2), delete/update is rejected PRIMARY KEY (sid,cid) CASCADE FOREIGN KEY (sid) also delete/update all REFERENCES Students tuples that refer to the deleted/updated tuple ON DELETE CASCADE SET NULL / SET DEFAULT ON UPDATE SET NULL) Set foreign key value of referencing tuple to NULL / given default 421B: Database Systems - The Relational Model 20 10 Where do ICs come from? ICs are based on the semantics of the real-world enterprise that is being described in the database relations We can check a database instance to see if an IC is violated, but we can NEVER infer that an IC is true by looking at an instance An IC is a statement about all possible instances For example, we might think that the student name is a key because in a given instance there do not exist two tuples with the same name; but it might happen in the future.