Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Open Database Connectivity wikipedia , lookup
Microsoft Jet Database Engine wikipedia , lookup
Extensible Storage Engine wikipedia , lookup
Concurrency control wikipedia , lookup
Entity–attribute–value model wikipedia , lookup
Clusterpoint wikipedia , lookup
Relational algebra wikipedia , lookup
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 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 53666 53688 53650 name Bartoli Chang Chang login faculty gpa bartoli@cs Science 3.4 chang@eecs Eng 3.2 chang@math Science 3.8 Column headers = attributes Cardinality = 3 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 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 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). VARCHAR(n) denotes a string of up to n characters (between 0 and n characters). SQL permits reasonable coercion between values of characterstring 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 Data Definition: Table Creation Defines all attributes of the relation The type/domain of each attribute is specified DBMS enforce correct type whenever a tuple is added or modified SQL is case insensitive It is possible to define default values Special NULL value: ‘unknown’ 421B: Database Systems - The Relational Model CREATE TABLE Students (sid CHAR(20), name VARCHAR2(20), login CHAR(10), faculty VARCHAR(20), gpa REAL DEFAULT 0.0) CREATE TABLE Enrolled (sid CHAR(20), cid CHAR(20), grade CHAR(2)) 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 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 Can update all tuples satisfying some condition DELETE FROM Students WHERE name = ‘Chang’ 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 53666 53688 53650 name Bartoli Chang Chang login faculty gpa bartoli@cs Science 3.4 chang@eecs Eng 3.2 chang@math Science 3.8 421B: Database Systems - The Relational Model name Bartoli Chang gpa 3.4 3.2 12 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 Primary and Candidate Keys in SQL Possibly many candidate keys exist, one of which is chosen as the primary key Each student has a unique id. For a given student and course, CREATE TABLE Students (sid CHAR(20) PRIMARY KEY, name VARCHAR2(20), … there is a single grade; further, no two students in a course receive CREATE TABLE Enrolled the same grade Application dependent Defined carelessly, an IC can prevent the storage of database instances that arise in practice! 421B: Database Systems - The Relational Model (sid CHAR(20), cid CHAR(20), grade CHAR(2), PRIMARY KEY (sid,cid), UNIQUE (cid,grade)) 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 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 Only students listed in the Students relation should be allowed to enroll for courses sid 53666 53688 53650 name Bartoli Chang Chang login CREATE TABLE Enrolled (sid CHAR(20), cid CHAR(20), grade CHAR(2), PRIMARY KEY (sid,cid) FOREIGN KEY (sid) REFERENCES Students) faculty gpa bartoli@cs Science 3.4 chang@eecs Eng 3.2 chang@math Science 3.8 421B: Database Systems - The Relational Model sid cid grade 53666 Topology112 C 53666 Reggae203 B 53650 Topology112 A 53668 History105 B 18 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 delete/update is rejected CASCADE also delete/update all tuples that refer to the deleted/updated tuple SET NULL / SET DEFAULT grade CHAR(2), PRIMARY KEY (sid,cid) FOREIGN KEY (sid) REFERENCES Students ON DELETE CASCADE ON UPDATE SET NULL) Set foreign key value of referencing tuple to NULL / given default 421B: Database Systems - The Relational Model 20 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. Key and foreign key ICs are the most common; more general ICs supported, too. (Discussed later in the course) 421B: Database Systems - The Relational Model 21 Logical Design: ER to Relational Entity sets to tables name eid salary CREATE TABLE Employees (eid CHAR(11), name VARCHAR(20), salary REAL, PRIMARY KEY (eid)) Employees Departments budget did Employees(eid, name, salary) Departments(did, dname, budget) CREATE TABLE Departments (did INTEGER, dname CHAR(20), budget REAL, PRIMARY KEY (did)) dname 421B: Database Systems - The Relational Model 22 Many-to-many Relationship Sets Map relationship set to table. Attributes of the table must include Keys for each participating entity set (as foreign keys) This set of attributes forms the key for the relation All descriptive attributes eid name salary Employees since Works_in Departments did dname budget 421B: Database Systems - The Relational Model Works_in(eid, did, since) CREATE TABLE Works_In (eid CHAR(11), did INTEGER, since DATE, PRIMARY KEY (eid,did), FOREIGN KEY (eid) REFERENCES Employees, FOREIGN KEY (did) REFERENCES Departments) 23 Relationships Sets with Key Constraints Alternative 1: map relationship set to table Many-one from entity set E1 to entity set E2: key of E1 i.e., key of entity-set with the key constraint is the key for the new relationship table (did is now the key) One-one: key of either entity set Separate tables for entity sets (Employees and Departments) eid name salary Employees Manages since Departments did dname budget 421B: Database Systems - The Relational Model Manages(eid, did, since) CREATE TABLE Manages (eid CHAR(11), did INTEGER, since DATE, PRIMARY KEY (did), FOREIGN KEY (eid) REFERENCES Employees, FOREIGN KEY (did) REFERENCES Departments) 24 Relationships Sets with Key Constraints (contd.) Alternative II: include relationship set in table of the entity set with the key constraint Possible because there is at most one relationship per entity Not useful if many entities do not have a relationship (wasted space, many not filled values) eid name salary Employees Manages since Departments did dname budget 421B: Database Systems - The Relational Model DepartmentsM(did, dname, budget, eidmgr, since) CREATE TABLE DepartmentsM (did INTEGER, dname CHAR(20), budget REAL, eidmgr CHAR(11), since DATE, PRIMARY KEY (did), FOREIGN KEY (eidmgr) REFERENCES EMPLOYEES(eid)) 25 Relationship Sets with Participation Constraints and Key Constraints Include relationship set in table of the entity set with the key constraint and the participation constraint We can capture participation constraints involving one entity set in a binary relationship if it also has a key constraint, but little else (at least within the table definitions) eid name salary Employees Manages since Departments did dname budget 421B: Database Systems - The Relational Model DepartmentsM(did, dname, budget, eidmgr, since) CREATE TABLE DepartmentsM (did INTEGER, dname CHAR(20), budget REAL, eidmgr CHAR(11) NOT NULL, since DATE, PRIMARY KEY (did), FOREIGN KEY (eidmgr) REFERENCES EMPLOYEES(eid)) 26 Renaming In the case the keys of the participating entity sets have the same names we must rename attributes accordingly name eid Reports_To(super-eid, sub-eid) salary CREATE TABLE Reports_To Employees (supervisor_eid CHAR(11), subordinate_eid CHAR(11), supervisor subordinate PRIMARY KEY (supervisor_eid, subordinate_eid), Reports_To FOREIGN KEY (supervisor_eid) REFERENCES Employees(eid), FOREIGN KEY (subordinate_eid) REFERENCES Employees(eid)) 421B: Database Systems - The Relational Model 27 Translating Weak Entity Sets Weak entity set and identifying relationship set are translated into a single table When the owner entity is deleted, all owned weak entities must also be deleted eid name salary Employees Policy cost Dependants pname dateofbirth 421B: Database Systems - The Relational Model Dependants_Policy(pname, dob, cost, eid) CREATE TABLE Dependants_Policy (pname CHAR(20), dob DATE, cost REAL, eid CHAR(11), PRIMARY KEY (pname,eid), FOREIGN KEY (eid) REFERENCES Employees, ON DELETE CASCADE) 28 Translating ISA Hierarchies General Approach: distribute information among relations Relation of superclass stores the general attributes and defines key Relations of subclasses have key of superclass and addit. attributes when a tuple in super-relation is deleted, corresponding tuples in sub-relation must be deleted Employees(eid, name, salary) Contract_Emps(contract_id,eid) CREATE TABLE Employees (eid CHAR(20), name salary eid name CHAR(20),salary REAL, PRIMARY KEY (ssn)) Employees CREATE TABLE Contract_Emps hours_worked (contract_id INTEGER, Contract_id hourly_wages eid CHAR(11), ISA PRIMARY KEY (eid), Hourly_Emps Contract_Emps FOREIGN KEY (eid) REFERENCES Employees, ON DELETE CASCADE) 421B: Database Systems - The Relational Model 29 Translating ISA Hierarchies (contd.) Object-oriented approach: Sub-classes have all attributes; if an entity is in a sub-class it does not appear in the super-class relation; Pro/Contra: Employees(eid, name, salary) Hourly_Emps(eid,name,salary, hourly_wages,hours_worked) Contract_Emps(eid, name, salary,contract_id) + A query asking for all hourly employees only has to go to one relation (in general approach it has to read two CREATE TABLE Employees relations) (eid CHAR(20) PRIMARY KEY, - Query on general attributes of all name CHAR(20),salary REAL) employees has to read all three tables - If an entity is both Hourly_emps and CREATE TABLE Hourly_Emps Contract_emps, name and salary are (eid CHAR(20) PRIMARY KEY, stored twice => undesired redundancy name CHAR(20),salary REAL, hourly_wages REAL, hours_worked REAL) CREATE TABLE Contract_Emps (eid CHAR(20) PRIMARY KEY, name CHAR(20),salary REAL, contract_id INTEGER) 421B: Database Systems - The Relational Model 30 Translating ISA Hierarchies (contd.) Last Alternative: one big relation Create only one relation for the root entity set with all attributes found anywhere in its network of subclasses. Put NULL in attributes not relevant to a given entity 421B: Database Systems - The Relational Model Employees(eid,name,salary, hourly_wages,hours_worked, contract_id) CREATE TABLE Employees (eid CHAR(20), name CHAR(20), salary REAL, hourly_wages REAL, hourly_worked REAL, contract_id INTEGER, PRIMARY KEY (eid)) 31 Translating Aggregation No key constraints Projects(pid,started_on,pbudget) Departments(did,dname,budget) Employees(eid,name,salary) Sponsors(pid,did,since) Monitors(pid,did,eid,until) Key constraint from Projects to Departments Projects(pid, started_on, pbudget, did, since) No Sponsors Monitors(pid, eid, until) Key constraint from Sponsorsname to Employees Sponsors(pid,did,eid,since,until) No Monitors Started_on pid pbudget Projects 421B: Database Systems - The Relational Model eid Employees salary Monitors until since Sponsors did dname budget Departments 32 Relational Model: Summary A tabular representation of data Simple and intuitive, currently the most widely used model. Integrity constraints can be specified based on application semantics (up to a certain degree). DBMS checks for violations Two important ICs : primary and foreign keys In addition, we always have domain constraints Powerful and natural query languages exist Rules to translate ER to relational model 421B: Database Systems - The Relational Model 33 Review: Binary vs. Ternary Putting any constraints in ssn upper picture: Key constraint on Policies name Employees (to guarantee that each policy only owned by one employee), would also mean that the policy can only cover one dependent Constraints of lower picture: Each dependant determined by one policy Each policy is owned by one employee covers pid Policies ssn name Dependants cost pname dob lot Employees Dependants beneficiary purchaser pid 421B: Database Systems - The Relational Model pname dob lot Policies cost 34 Binary vs. Ternary Key constraints allow us to combine Purchaser with Policies and Beneficiary with Dependants Participation constraints lead to NOT NULL constraints (or primary key in case of weak entity) What if policy is a weak entity? 421B: Database Systems - The Relational Model Policies(pid, cost, eid) Dependants(pname, dob, pid) CREATE TABLE Policies (pid INTEGER, cost REAL, eid CHAR(11) NOT NULL, PRIMARY KEY (pid), FOREIGN KEY (eid) REFERENCES Employees ON DELETE CASCADE) CREATE TABLE Dependants (pname CHAR(20), dob DATE, pid INTEGER, PRIMARY KEY (pname,pid), FOREIGN KEY (pid) REFERENCES Policies, ON DELETE CASCADE) 35