Summary: Possible simplifications in query processing are investigated if, consciously, it is agreed to use query languages of limited power (i.e., weaker and proper sublanguages of first-order language). In such a situation, it is frequently possible to transform data in the database into some other form that is not necessarily logically equivalent to the original data but that preserves all the answers for the queries from our query sublanguage. This transformation is beneficial if it leads to a database that is easier to deal with. Such a technique is called an abstraction. It is demonstrated that a well-known chase transformation is an abstraction that preserves information modulo positive queries (i.e., queries without negation) when applied to the tables with null values (\(V\)-tables). On the other hand, it is shown that no similar abstraction can be constructed for restricted tables in which marked null values vary only through some finite subsets of domain. It is also shown a number of other types of abstractions and their applications.