RapidMiner's most important characteristic is the ability to nest operator chains and build complex operator trees. In order to support this characteristic, the RapidMiner (formerly YALE) data core acts like a data base management system and provides a multi-layered data view concept on a central data table which underlies all views.

For example, the first view can select a subset of examples and the second view can select a subset of features. The result is a single view which reflects both views. Other views can create new attributes or filter the data on the fly. The number of layered views is not limited.

This multi-layered view concept is also an efficient way to store different views on the same data table. This is especially important for automatic data preprocessing tasks like feature generation or selection. For example, the population of an evolutionary operator might consist of several data views - instead of several copies of parts of the data set.