The DataSet is one of the more widely used components of ADO.NET. It is a key element of the disconnected programming model that ADO.NET is based on, and it enables you to explicitly cache data from different data sources. For the presentation tier, the DataSet is tightly integrated with GUI controls for data-binding. For the middle-tier, it provides a cache that preserves the relational shape of data, and includes fast simple query and hierarchy navigation services. A common technique used to lower the number of requests on a database is to use the DataSet for caching in the middle-tier. For example, consider a data-driven ASP.NET Web application. Often, a significant portion of the application data does not change frequently and is common across sessions or users. This data can be kept in memory on the Web server, which reduces the number of requests against the database and speeds up the user’s interactions. Another useful aspect of the DataSet is that it allows an application to bring subsets of data from one or more data source into the application space. The application can then manipulate the data in-memory, while retaining its relational shape.

Despite its prominence, the DataSet has limited query capabilities. The Select method can be used for filtering and sorting, and the GetChildRows and GetParentRow methods can be used for hierarchy navigation. For anything more complex, however, the developer must write a custom query. This can result in applications that perform poorly and are difficult to maintain.

LINQ to DataSet makes it easier and faster to query over data cached in a DataSet object. These queries are expressed in the programming language itself, rather than as string literals embedded in the application code. This means that developers do not have to learn a separate query language. Additionally, LINQ to DataSet enables Visual Studio developers to work more productively, because the Visual Studio IDE provides compile-time syntax checking, static typing, and IntelliSense support for LINQ. LINQ to DataSet can also be used to query over data that has been consolidated from one or more data sources. This enables many scenarios that require flexibility in how data is represented and handled. In particular, generic reporting, analysis, and business intelligence applications require this method of manipulation.