Database for managing the databases

Databases collect data and link them into logical units. The individual data are provided with information and meta descriptions that are needed for their processing. A database can be used for managing databases and to facilitate the retrieval of specific information. In many databases, rights can be defined that determine which programmes or persons have access to which data. Most databases can only be opened, edited and read by special database applications.

Only valid queries are made in the correct order, thus completing the whole transaction correctly. Successful transactions leave a stable database and require constant checking. Transactions do not "get in each other's way", which is usually ensured with the help of certain locking functions. All data is stored permanently, even after a successful transaction has been completed. This is especially true in the event of failures or system errors. Essential for this are transaction logs, for example, which record all processes.

The hierarchical database identifies unique dependencies so that, except for the root, every record has exactly one predecessor. Levels that are not directly adjacent cannot interact with each other. Hierarchical databases are very clear, but extremely inflexible.

Network database models do not have strict parent-child relationships. All data sets can have several predecessors, which results in a network-like structure. Thus, there is no unique access path to a record. Data records can be inserted and removed fluidly without interfering significantly with the overall structures. Network-like database models are mainly used on mainframe computers. Customers of IBM trust the hierarchical model. A well-known network model is the UDS from Siemens.

In the relational database, SQL is usually used as the database language. The formulation of relations is done by relational algebra, with which specific information can be obtained from relations. The principle in turn is the basis of the database language SQL. Individual tables define the localisation of and links between the information. A data set is formed from this information. Individual pieces of information are collected as attributes in the columns. The overall relation results from the interrelated attributes. The primary key, which is usually defined as the first attribute that must not change, is used to uniquely identify data records.

Object-oriented databases are partly available as open source and are most commonly used on Java and .NET platforms. The data is stored together with its access methods. Methods or functions are stored in an object. With the help of the "methods", access to these objects is defined, which are stored together with the data in the object. The objects can be complex and consist of any number of types of data. They also receive a unique identification number. Individual objects are grouped into object classes, resulting in a class hierarchy. New, complex objects can be fed in relatively simply and smoothly.

In the document-oriented database model, documents are the basic unit for storing data. The data is stored in key/value pairs. All documents are self-contained units. A single document is sufficient to store all information. Related data is always stored together in one place. Targeted querying of a document is sufficient. Document-oriented database systems are primarily of interest for web applications, as complete HTML forms can be fed in. However, not all document databases are suitable for all areas of application.