Statistical matching procedures are methods that aim at the integration of sources. A description of real contexts in the Italian National Statistical Institute, in which these methods appear to be useful, is given. The necessary steps for the application of statistical matching are analysed. Special emphasis is given to statistical methods available in literature. The key role of auxiliary information in matching process is highlighted. An ad hoc example shows how logical constraints represent an important source of auxiliary information in the context of Official Statistics. In our example the introduction of this particular information in the matching step is achieved through the EM algorithm. Finally, a number of open problems are discussed.