Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Next revision		Previous revision
protogrid:database_design_fundamentals [2017-11-15 19:09] – created gawprotogrid:database_design_fundamentals [2018-03-26 07:49] (current) – [Database Design Fundamentals] 77.58.53.50
Line 1: Line 1:
 ====== Database Design Fundamentals ====== ====== Database Design Fundamentals ======
  
-The real world is full of objects: +Database design is an important task for each software developer. It is in most software development projects crucial to understand the real world problem thoroughly and to be able to design the most simple database design possible. The key to success is as well a good understanding of the problem given as the ability to simplify the real world to a good model. The most talented designers master the art of designing a data model that is universal and versatile and continues to work unchanged when there will be added entities and attributes in the future.
-* physical and tangible objects like houses, dogs, cars etc. +
-* virtual objects like orders, requests, code libraries etc.+
  
-We call each class of objects entity.+==== Entities ==== 
 +The real world is full of objects. If you look around you and think a bit you will find 
 +  - //physical// and tangible objects like houses, dogs, cars etc. 
 +  - //virtual// objects like orders, requests, code libraries, international organizations, countries etc.
  
-There are relations between entities: +We call each class of objects of the same type //entity// (from latin ens //being//). 
-* mother-child-relations: there might be many, one or no child, we call this a 1-to-many-relation and write 1:m + 
-* child-mother-relation: there is only one mother for each child, we call this a m-to-1-relation and write m:1 +An entity is usually represented in software by a form. In table-oriented databases, an entity can also very easily be understood and represented as a table (consisting of rows and columns). In Protogrid, each entity is defined as a //[[Proto]]//, a special Design Card that is used to define the attributes, appearance and behaviour of Data Cards. 
-* woman-mother-relation: each woman might be a mother or not, we call this a 1-to-c-relation and write 1:c+ 
 +==== Records = Tuples = Rows = Data Cards ==== 
 +Entity records have different names in literature. In table-oriented databases it is represented as a table row. More generally it is called a tuple, i.e. a set of attributes defined in an entity. In Protogrid, records are represented as //[[protogrid:card|Data Cards]]//. An entity usually has many records. Each record has a unique key that allows to identify the record. In Protogrid, that key is called a //Card Key//. 
 + 
 +==== Attributes = Fields = Columns ==== 
 +An entity has attributes. Each attribute describes a part of the entity. In table-oriented databases, an attribute is represented as a table column. In Protogrid, attributes are represented as //[[protogrid:field|fields]]// on a Data Card, and are defined using //Field Definitions// on the Proto. Field Definitions are another type of Design Card, responsible to define the behaviour and appearance of fields. 
 + 
 +As an example, a house might have the following attributes: 
 +  * its address 
 +  * the number of floors 
 +  * the type of roof 
 +  * the color of the front 
 +  * the number of parking lots 
 +  * etc. 
 + 
 +An attribute is typically of a specific data type, like: 
 +  * text field 
 +  * number field 
 +  * date 
 +  * time 
 +  * date and time 
 +  * etc. 
 + 
 +==== Relations ==== 
 +Often there are **relations** between entities. 
 + 
 +=== The 1:n-relation (Mother-Child Relations) === 
 +For one mother there might be many, one or no child. We call this a 1-to-many-relation and write 1:n, with //n// representing the variable number of children. //1// and //n// are often called the //cardinality// of an entity in a relationship. In database design, the variable //n// is usually repeated for each of the relations, even though each represents the cardinality of a different entity. 
 + 
 +Many times we see mother-child-relations: 
 +  * an organization has members, hence  
 +    * there is a 1:n-relation between the organization and its members 
 +    * the organization has (no, one or usually many) members 
 +    * a member belongs to an organization 
 +  * an organization has different office locations 
 +    * there is a 1:n-relation between the organization and its office locations 
 +    * the organization has (no, one or usually) many office locations 
 +    * an office location belongs to exactly one organization 
 +  * a house has floors 
 +  * a floor has rooms 
 +  * a room has doors 
 + 
 +How do we represent these relations in a software's user interface? 
 +  * the mother entity is a form with fields 
 +  * the child entity is also a form with fields 
 +  * on the mother form we have an embedded view showing all children 
 +  * on the child form we have a field that shows us the name of the mother and a button or link to go to the mother form 
 + 
 +This knowledge is already enough to build many different types of applications. 
 + 
 +== How to implement 1:n-relations in Protogrid == 
 +  - create the [[Proto]] for the child entity 
 +  - create the Proto for the mother entity 
 +  - add a [[TableView]] to the mother Proto which relates to the child Proto 
 +  - add a relation [[field]] to the child Proto which relates to the mother Proto 
 + 
 +That's all folks! It really is that simple in Protogrid. This is where Protogrid truly shows its Rapid Application Development capabilities. 
 + 
 +There is one more important variant of relations: 
 + 
 +=== The n:m-relation === 
 + 
 +Sometimes a child has not only one related entity (like a mother - there is only one for each child on this planet) but many. The closest example to think of are siblings. 
 + 
 +Examples: 
 +  - an order might include several articles and an article might be part of several orders 
 +  - a person might be member of several organizations and an organization has usually several members 
 + 
 +For the sake of simplicity any n:m-relation might be split into two 1:n-relations 
 + 
 +Examples: 
 +  - there are the three entities order, article and article-in-order 
 +    * the order has as attributes: order number, order date 
 +    * the article has as attributes: article number, article description, price 
 +    * the article-in-order has as attributes: the order number, the article number - just two attributes that is usually enough 
 +  - there are the three entities organization, person and member 
 +    * the organization has as attributes: organization ID, name 
 +    * the person has as attributes: person ID, name, age 
 +    * the member has as attributes: organization ID, person ID - just two attributes 
 + 
 +== How to create an n:m-relation in Protogrid == 
 +  - create the first [[Proto]] (e.g. the organization) 
 +  - create the second Proto (e.g. the person) 
 +  - create the relation Proto with only two fields of type relation, pointing to the first and the second Proto, respectively. 
 +  - go to the first Proto and add a [[TableView]] with the relation Proto 
 +  - go to the second Proto and add a TableView with the relation Proto 
 + 
 +In Cards of the first Proto you now can see all relations to the other Proto. 
 +And in Cards of the second Proto you can see all relations to the first Proto, which is exactly what we need in these cases. 
 + 
 + 
 +==== Summary ==== 
 +  * Entity ~ form ∼ table ~ Proto 
 +  * Record ~ row in a table ~ Data Card 
 +  * Attribute ~ column in a table ~ field

Trace:

Print/export