Vous êtes sur la page 1sur 2

In a very generic sense, the two may appear to be similar, but there are very important differences, in architecture,

technology and usage patterns. The term "database" is very broadly defined and covers a lot of stuff, but the basic ideas are relatively simple to understand. A database, in the more common meaning of the word, is repository of information that is used as a backing storage for some specific application or set of applications. Databases are usually structured, but this is not a definitive technical requirement. Databases are often used in contexts where you need to be able to access and update data online, in near real time, and with multiple concurrent accesses. Real world, professional-use databases have stricter requirements; for example, security, and also consistency, in the form of transactions, that guarantee that the information can be reliably used by their clients. A data warehouse is a particular type of database, which focus on a very specific application: storing, filtering, retrieving and analyzing huge volumes of information. This application imposes a different set of constraints and leads to a completely different architecture and usage pattern. 1. Inside the same company, data warehouses are potentially much bigger than the databases from where the data is derived. Data warehouses have to store much older historical records, and sometimes have to store also information that is bought or captured elsewhere to complement the information that is generated and stored by the internal database system. 2. Access patterns are different. Database records are often read one by one; data warehouses are frequently acessed by reporting engines that have to process entire datasets, or columns, at a time, to generate aggregates and other analytical information. 3. Very often in a data warehouse you have to store different versions of the same basic data (that comes from tables in a standard database). In a traditional database you would probably need to migrate all the data, in a data warehouse, you just compose you reports using data from multiple sources. 4. Data warehouses are structured in a completely different format

regarding its internal relationships. There is much less focus on traditional indexes and joins, and relationships are expressed by other mechanisms. You just don't do a SQL query with a join, you (usually) use a custom query builder to create your reports (yes, you don't call them queries), using features that are specific for your data warehouse provider and information architecture

Vous aimerez peut-être aussi