Presentation on theme: "Normalization Rules for Database Tables Northern Arizona University College of Business Administration."— Presentation transcript:
Normalization Rules for Database Tables Northern Arizona University College of Business Administration
Transforming EER Diagrams into Relations (cont.) Mapping Supertype/Subtype Relationships –One relation for supertype and for each subtype –Supertype attributes (including identifier and subtype discriminator) go into supertype relation –Subtype attributes go into each subtype; primary key of supertype relation also becomes primary key of subtype relation –1:1 relationship established between supertype and each subtype, with supertype as primary table 2
Supertype/Subtype Relationship in ER Studio Notation One to One Relationships, X Indicates completeness, No Notation for disjoint vs.overlapping Primary key of supertype becomes both Primary and Foreign Key In subtypes
5 Figure 4-21 Mapping supertype/subtype relationships to relations These are implemented as one-to-one relationships
Normalization - Some Definitions a relation is a two-dimensional array with a single-valued entry in each cell which has no duplicate rows and has columns whose meaning is the same across all rows. –All tables used in the relational model must be relations. Normalization is a process for evaluating table structure and reorganizing them as needed to product a set of stable, well-structured relations. An anomaly is a condition which interferes with the storage, or retention of data or creates the potential for inconsistent data. –There are insertion, modification, and deletion anomalies. –The Normalization process should eliminate anomalies.
Unnormalized Tables An Unnormalized table is a table that does not meet the definition of a relation. –it contains rows with multiple values for an attribute (repeating groups) or –contains duplicate rows. A table is said to be in first normal form if it meets the definition of a relation –Generally this means it contains no repeating groups of attributes. The next slide shows an example of an unnormalized table.
EMPLOYEE This EMPLOYEE table is unnormalized - It has cells that do not contain single-valued entries. As shown this table has no logical primary key. The E ID# does not functionally determine the value of Skill.
EMPLOYEE The above employee table shows the same set of data as the previous slide. It has been reorganized into a form that could be implemented under some file processing systems, using COBOL, for instance. However, it is still not in a form that can be used by the relational model. The Skills are a multi-valued (repeating group) of attributes which cannot be identified by the primary key.
Eliminating Repeating Groups In most cases, Unnormalized tables can be converted to Sets of Tables that are in at least First Normal form by: Original Table EMPLOYEE placing any repeating groups of fields in a separate table which includes the primary key attribute from the original table along with a single occurrence of the repeating attribute (Skill in our example). A Table is in first normal form if it contains no multi-valued attributes
Eliminating Repeating Groups EMPLOYEE Original Normalized EMPLOYEE_SKILL
Logical ER Diagram in ER Studio Notation 1 st Normal Form Example
This Table is in First Normal Form Schedule of Classes There are no repeating groups of attributes. NOTE: The primary key of this table is a Concatenated key - no single attribute uniquely identifies a row of the table, but the Combination of Course # and Section # does uniquely identify a row. If I know that the Course is CIS 120 and the Section is section 1, I can identify a unique schedule occurrence.
Although this Table is in First Normal Form, it Contains Anomalies Schedule of Classes If the description of CIS220 Changes from VB Prog. to Visual C#, I must record the new value in two places (as shown) - This is a modification anomaly
Although this Table is in First Normal Form, it Contains Anomalies Schedule of Classes If a new course has been designed and I know its description and credit hours (ACC 266, Pers. Acc., 2 hrs), I still cannot record this data until at least one section of the course is offered - an insertion anomaly. ?
A Table in First Normal Form Containing Anomalies Schedule of Classes If no section of ACC 255 is offered this semester, I will lose the information about the description and credit hours of this course. - A deletion anomaly
Schedule of Classes This table has anomalies because it contains partial dependencies. A partial dependency occurs when one or more attributes in a table depends upon (is functionally determined by) only a portion of a concatenated primary key. In this case the Description and Cr. Hrs. attributes depend only on Course #. To correct this problem, those attributes determined by only a part of the key should be placed in a separate table. Its Primary key will be the portion of the original primary key required to identify them.
Schedule of Classes Original Schedule of Classes COURSE Revised Notice how this structure eliminates the anomalies we found
Logical ER Diagram in ER Studio Notation 2 nd Normal Form Example
Second Normal Form Partial Dependencies occur when nonkey attributes are functionally determined by only a portion of a concatenated primary key. Partial dependencies can occur only in tables with a concatenated key. Partial dependencies can be corrected by removing those attributes to a separate table whose primary key is just the portion of the key from the original table needed to functionally determine them. A table is in second Normal Form if it is in first normal form and it contains no partial dependencies.
A Table in Second Normal Form Which Has Anomalies PROFESSOR This table is in 2nd normal form since it has no repeating groups of attributes (first normal form) and its primary key is not concatenated. However, the table above still has anomalies.
Anomalies in the Example Professor Table PROFESSOR Modification Anomalies -if the Dept Aide serving the ECO department changes, or if the Fax # of the ECO department changes, this change would need to be made in several records. Insertion Anomalies - I want to start a new department and have a Dept Code, a Dept Aide, and a Dept Fax # (e.g., MKT, T. Taylor, 523-7216). I can’t add this data to the table until at least one professor is hired to teach in this new department.
Anomalies in the Example Professor Table PROFESSOR Deletion Anomalies - If Prof # L29 (the only professor in the CIS department in our example table) is deleted, we would lose the information about the name of the Dept Aide for CIS and the Dept Fax # for CIS.
This Professor Table has Transitive Dependencies PROFESSOR The anomalies we have found occur because the Professor table has transitive dependencies. Dept Code, Dept Aide, and Dept Fax # are all attributes of a DEPARTMENT entity which is uniquely identified by Dept Code - If I know Dept Code I can uniquely identify Dept Aide and Dept Fax #. Knowing Prof # allows me to identify these attributes, but only through a chain of inferences - Prof # uniquely identifies Dept Code which, in turn uniquely identifies the other DEPARTMENT attributes. The anomalies can be resolved by removing the attributes determined by a non-key attribute to a separate table.
Correcting Transitive Dependencies PROFESSOR Original PROFESSOR DEPARTMENT Revised
Logical ER Diagram in ER Studio Notation 3rd Normal Form Example
Third Normal Form Transitive dependencies occur when non-key attributes are functionally determined by other non-key attributes. Transitive dependencies can be corrected by removing the attributes to a separate table whose primary key is the attribute of the original table which functionally determines them. The functionally determining attribute serves as a foreign key in the original table. A table is in Third Normal Form if it is in second normal form and it contains no transitive dependencies.
A Proposed Normalization Process for Database Designers Examine each table of the proposed structure and perform the following operations: Remove any repeating groups of attributes (multi-valued attributes) to a separate table. If there are independent sets of multi-valued attributes place each set in a separate table. Remove any attributes that are functionally determined by only a portion of a concatenated key to a separate table. Remove any attributes that are functionally determined by a non-key attribute to a separate table.
Review Question: What Normalization rule(s) are violated by the table below? How would you revise the table Structure? Write out your answer on a piece of scratch paper.
Review Question Solution: Original Table: This table violates both 2nd & 3 rd normal forms. Emp. Name and Emp. Class both depend only on Emp No. Which is part of the concatenated key - violates 2nd normal form. Wage Rate is actually determined by Emp Class a non-key attribute which violates 3rd normal form. Normalized Tables Employee Class Employee Employee Hours
Logical ER Diagram in ER Studio Notation Review Question Example
32 b) Resulting relations in ER Studio Notation Example of mapping a binary 1:1 relationship (cont.) Foreign key goes Into the Entity on the optional side of the relationship
Mapping Unary Relationships One-to-Many–Recursive foreign key in the same relationOne-to-Many–Recursive foreign key in the same relation Many-to-Many–Two relations:Many-to-Many–Two relations: –One for the entity type –One for an associative relation in which the primary key has two attributes, both taken from the primary key of the entity 33
34 Figure 4-17 Mapping a unary 1:N relationship (a) EMPLOYEE entity with unary relationship (b) EMPLOYEE relation with recursive foreign key ER Studio Notation
35 Figure 4-17 Mapping a unary 1:N relationship (a) EMPLOYEE entity with unary relationship (b) EMPLOYEE relation with recursive foreign key
36 Figure 4-18 Mapping a unary M:N relationship (a) Bill-of-materials relationships (M:N) (b) ITEM and COMPONENT relations ER Studio Notation with Sample data ITEMCOMPONENT ItemNoItemDescripItemNoComponentNoQuantity ADC8Audio CardPCD2ADC81 MBD2MotherboardPCD2MBD21 PCD2PC Dual CorePCD2RAM94 PCQ5PC Quad Core... RAM91GB RAM ChipPCQ5ADC81... PCQ5MBD21 PCQ5RAM98...
37 Figure 4-18 Mapping a unary M:N relationship (a) Bill-of-materials relationships (M:N) (b) ITEM and COMPONENT relations
Merging Relations View Integration–Combining entities from multiple ER models into common relationsView Integration–Combining entities from multiple ER models into common relations Issues to watch out for when merging entities from different ER models:Issues to watch out for when merging entities from different ER models: –Synonyms–two or more attributes with different names but same meaning –Homonyms–attributes with same name but different meanings –Transitive dependencies–even if relations are in 3NF prior to merging, they may not be after merging –Supertype/subtype relationships–may be hidden prior to merging 38
Enterprise Keys Primary keys that are unique in the whole database, not just within a single relationPrimary keys that are unique in the whole database, not just within a single relation Corresponds with the concept of an object ID in object-oriented systemsCorresponds with the concept of an object ID in object-oriented systems 39
40 Figure 4-31 Enterprise keys a) Relations with enterprise key b) Sample data with enterprise key