Data integration is a process in which the data are combined. In data integration process the data are combined and then it is provided to the user in a unified view. Data integration process can be finding in both ways either commercial or scientific. The need of sharing the existing data explodes had appeared the data integration process. There are numerous open problems still remain to be solved. Sometimes data integration is also called Enterprise Information Integration in management practice.

Data integrations had emerged due to the problem of combining heterogeneous data sources under a single query interface. And after the 1960 the need to share existing repositories had was the reason behind the rapid adoption of databases. At several levels these merging of data can be done in database architecture. Among these the most popular approach is Data Warehousing. In this the data are transformed, extracted and loaded into different source and from there it can be queried by a single schema because the data reside together within a single repository at the time of query. Sometime the problem regarding tight coupling arises.

Constructing a data warehouses when there is a query interface to the data sources and there is no access to full data then it is very difficult. This type of problem arises during the integration of commercial query services which includes travel or classified advertisement web applications.

The subset of database theory which formalizes the underlying concepts of the problem in first-order logic is called the theory of data integration. The result of this tells us how data integration is difficult to perform. The definition of data integration is very simple but the work is not as simple as the definition.

The systems of data integration is defined as G,S,M, in this G means the global schema, S means the heterogeneous set of source schemas and M is the mapping which is used to maps queries between the source and the global schemas. Here G and S are generally expressed in form of alphabets which includes symbols for every of their relations. Between the queries over G and queries over S, the mapping M consists of assertions. In this system when the user pose queries over the data integrations system, then they pose queries over G and at the same time the mapping then asserts the connections between the elements in the global schema and the source schemas.

Query processing is also very important part of data integration. In data integration systems the theory of query processing is defined using conjunctive queries. There are some formal languages also which express these query very concisely and also without ambiguity. Formal language like Datalog. Most of the time SQL queries are also classified as conjunctive queries as well. Query containment is an important part of query processing. In this if there is two queries then the two queries are said equivalent if the resulting sets are equal for any database.

The commercial application of Data Integration is called Enterprise Information Integration (EII). Here the private sector is more concerned regarding the problems of data integration. The are several problems related to EII such as it must be simple to understand, must be simple to employ, must handle higher order information.