Abstract
The schema integration technique offers the possibility to unify the representation of several schemas into one global schema. In this work, we present two contributions. The first one is about automating this technique to reduce human intervention. The second one is about applying this technique to generate data warehouse schema from data mart schemas. To response to our goals, we propose a new methodology that is composed by schema matching and schema mapping. The first technique compares the elements of the two schemas using a new semantic measure to generate the mapping rules. The second one transforms the mapping rules into queries and applies them to ensure the automatic merging of the schemas.