System Requirements Of The Applications

Published Date: 02 Nov 2017

Example of a Database

A good database design requires an understanding the functions we want to model and the database features and concepts to represent that business functions. To design a perfect database, it is important to model the business, because it might change the design of database once implemented moreover it can be time consuming. In order to perform better, a well-designed database is required. While we designing a good database, we might need to take note on:

Create a database plan to fit the database and how it affects the design.

Use normalization rules to prevent mistakes in the database design.

Protect our data integrity

User permissions and security requirements of the database

System Requirements of the applications.

Example of Simple Database Design:

So how we interact with the database? Form and reports will do the way. Forms allow us to key in the new records and edit existing records. Reports allow us to show exactly what we want to see by sorting and grouping the results. Reports can show totals for the whole reports of for groups. Example: Avg, Count, Sum, Min, Max, etc.

A perfect designed form is easy to read and use. A data entry forms should be matches the information that the users enter a new data. Reading the data should not interfere by background. Entered value should be clearly different from calculated value.

A perfect design report is printed neatly and easy to read. Totals and grouping should be clearly labeled. The control sources show all of the data and the reports fit on the page.

Example of Form:

Example of Report:

Example of report:A poor database structure can be cause due to lack of experience, lack of time, shortage of the necessary skills and insufficient resources. Besides, a poor database design can lead to lots of problem down the line, such as performance, unable to accommodate new features and low quality of data.

Documentation for database usually divided into three parts: inaccurate, incomplete, or none at all. It never centralized in one place if it would make it worse. It would be easier for us if we could have a proper, centralized documentation.

Denormalize a database will lead to sacrificing flexibility but will obtain optimized performance.

A special application code is needed to handle because there will be lots of NULLs for columns that don’t apply to a record although the "One table" approach may make data access easier.

Another well-known mistake is repeating values.

That can greatly decrease flexibility and increase difficulty when updating the data.

Example of Denormalized structure:

Example of Normalized equivalent:

The first form inform us to break off Phone and Address information into their own tables with identifiers.

The second normal form, break off the Zip Code information into the table so that it is not duplicated.

Consideration recommended to be made for City and State information, but that might be a performance over duplication wins.

The third normal form, look at the Industry in the Customer table and Organization Name and put those into the tables since they are not functionally on the Customer ID. This would apply to City and State Code in the ZIP Code table.

A poor database design will lead to inefficient storage, conflicts in data and difficulties in maintenance. It’s true that everyone will make mistakes and end up with a poor database design, but of course not all of them. Therefore, what are the common mistakes that we always done?

First, I always see data types chosen as if they don't matter.

If we know that the possible values columns are between 0 and 100,000 then we don't need to type a BIGINT for the column as an INT would do just fine.

The BIGINT data type requires 8 bytes of space, and the INT need only 4 bytes of space. It not only save our time but save our storage too.

I show you an example for one column, but what about thousands of columns that are storing first and last names? How much extra for those? How about your date columns? Do you need a calendar dates prior to the year 1910, or after the year 2080? If not, SMALLDATETIME will be work just fine for us.

Next, If we are using a surrogate key in our database design, we should not forget that the data in our table possible have an alternate key based on the business logic that was used to the creation of the table.

Many database designers forget all about the other columns and tend to focus on the surrogate key only.

This is bad for our data quality as well, it would allow for duplicate values into our business layer if we do not put an index on that alternate key.

Example of poor database design:

Of course everyone will accidentally make some mistake while designing a database, but how could we avoid it even a small mistakes? One of the best ways is to organize data. A custom database application can be built for us to meet our requirement needs. Both Professional and practical business needs are one of the best database systems.

Never create a poor design. This might seem obvious, however if the designer does not know where he/she is going, he/she could prefer to use any ways to create a database.

Whenever we create a database, we might want to establish one that gets you what you want while making sure the database won't fail our application down.

In order to create a good database, we need to first create a perfect plan.

Plan out the needs of our project and how our database is going to meet those requirements.