Target
In this table we underline the identification data. The source of each PTH must be the identification key data, the target may not be the identification key data. When the target is not the identification key then we have an entity (in the case of entity A:).
In case the identification key is D, the target is A, the identification key of entity A, we detect a match between entity D and entity A through the hierarchical association: D---A.
SKIN
Similarly we detect the functional dependency between C and D by hierarchically combining C – D.
D | 1.1 / 0.1 |
D | |
Maybe you are interested!
-
Solutions for developing application information systems in the current Vietnamese banking system - 17 -
Information and Communication Systems Communication Systems -
Research on factors affecting the quality of accounting information systems in Vietnamese traffic construction enterprises - 26 -
Master's Thesis Information Systems: Investment Project Evaluation and Automatic Project Management Scheduling - 1 -
Information Systems - 61
1,N / 0,N | A |
A B | |
After considering the columns and rows, there may exist data that are not the target of any PTH whose source is the considered data. In our case, it is data E. Let us consider the unions of primary data that may be the source of a PTH whose target is data E. Here, for example, it is A + C. Then we have a non-hierarchical association between the corresponding entities (A, C).
We can add a column that allows the
This PTH is presented because there is a new source:
Source | ||||
1 | 3 | 4 | 1 + 3 | |
1. A | 1 | 1 | 1 | |
2. B | 1 | |||
3. C | 1 | 1 | ||
4. D | 1 | 1 | ||
5. E | 1 | |||
Target
We can also add a row if we want.
shows the reflex of PTH.
Note :
To be able to detect functional dependencies where the source is a combination of single sources (in our case A, C, D) we have to consider combinations , where m is the number of functionally dependent source data, n is the number of data considered. In our case m = 3, n = 5 and the possible combinations are (A, C; A, D; D,
C) and (A, C, D). According to the hypothesis A, C E do
So, we do not need to consider other combinations.
A data must be represented in an entity or an association. Therefore, on its corresponding row there must be at least one 1.
However, there are two cases that need to be considered. Data that is not identification data may in fact have more than one "1" on the corresponding line.
In this case, it proves the existence of a PTH.
It is necessary to remove these PTHs.
Identification data can have multiple "1"s on the corresponding row. Some "1"s specify the identity (the digit is in the cell with the same row order as the column order), other "1"s specify the existence of hierarchical combinations such as the "1" in row 1, column 4, or non-hierarchical combinations such as row 1,3 column "1+3".
Data that is neither source nor destination of a PTH is marked by writing "parameter" on the corresponding line.
Conclude :
An entity is characterized by an identifying attribute and other attributes that represent the nature of the entity, which is the target of a PTH whose source is the identifying data.
For example: entity A, identity attribute is A_ and regular attribute is B. Two types of associations are defined: