Generalized joins
Generalized joins
复制标题
广义连接
DOI:
--
复制
发表时间:
1976
期刊:
影响因子:
--
通讯作者:
A. Pirotte
中科院分区:
文献类型:
--
作者:
M. Lacroix;A. Pirotte
In [I ], Codd proposes a three-valued logic to describe the behavior of null values in the usual operations of a relational algebra. This short note presents generalizations of the join operation which are defined in such a manner that no information is lost in the join, i.e. that the operand relations can be recovered from the join. The proposed operations introduce null values (denoted by "~") in the join of two relations which are supposed to contain no null value. If they do contain m, then the joins are easily modified by applying the rules given in [11. The definition of the generalized joins are given below for the join of R(A, BI) with S(B2, C), where BI and B2 participate in the join. Modifying the number of domains of the relations is simply a matter of notations. The following extensions will be considered as an example : R S R [B1 ~ B2 ]S A B1 B2 C A B1 B2 S u m n a u p p b u ~ p b w n n a w n q c x p p b x p u m m x r x r m ~ q c The generalized equi-join R [ BI ~ B2 ]S is shown above for the example. U {~,~>} ® (s*(s [ 1 ]-R [2 ])).