Citation Context

...rocessing queries on U-relations. In our experiments we show that query evaluation on U-relations scales to large amounts of data with high degrees of uncertainty. 1 Introduction Several recent works =-=[10, 9, 8, 2, 14, 4, 6]-=- aim at developing scalable representation systems and query processing techniques for large collections of uncertain data as they arise in data cleaning, Web data management, and scientific databases...

Citation Context

... tend to differ in succinctness, several have polynomial-time data complexity for (decision) problems such as tuple possibility under positive (but not full) relational algebra. This includes vtables =-=[12, 11]-=-, uncertainty-lineage databases (ULDBs) [8], and world-set decompositions (WSDs) [6]. 4. Ease of use for developers and researchers in the sense that the representation system can be easily put on top...

Citation Context

...rocessing queries on U-relations. In our experiments we show that query evaluation on U-relations scales to large amounts of data with high degrees of uncertainty. 1 Introduction Several recent works =-=[10, 9, 8, 2, 14, 4, 6]-=- aim at developing scalable representation systems and query processing techniques for large collections of uncertain data as they arise in data cleaning, Web data management, and scientific databases...

Citation Context

...delled using three variables x, y, z that each can independently take one of two values. We model this scenario by the U-relational database shown in Figure 1b. We use vertical partitioning (cf. e.g. =-=[7, 15]-=-) to achieve attribute-level representation. R is represented using four U-relations, one for each column of R. The U-relation for the coordinate positions (which are all certain) is not shown since w...

Citation Context

...mula (x 1≤i≤k i 1 = vi 1 ∧ ... ∧ xi mi = vi mi ). Recent work considered efficient solutions for restricted classes of queries and probabilistic databases [10] or by applying approximation techniques =-=[14]-=-. Scalable confidence computation is out of the scope of this paper. Our current approach for exact confidence computation exploits the independence and variable sharing among ws-descriptors and is by...

Citation Context

...uation. A trade-off is required between the succinctness of a representation formalism and the complexity of evaluating interesting queries. This tradeoff follows from established theoretical results =-=[1, 11, 6]-=-. Saarland University Database Group Saarbrücken, Germany {lublena, jansen, koch, olteanu}@infosys.uni-sb.de 1 However, while the formalisms in the literature tend to differ in succinctness, several h...

Citation Context

...-relations. In our experiments we show that query evaluation on U-relations scales to large amounts of data with high degrees of uncertainty. I. INTRODUCTION Several recent works [10], [9], [8], [2], =-=[15]-=-, [4], [6] aim at developing scalable representation systems and query processing techniques for large collections of uncertain data as they arise in data cleaning, Web data management, and scientific...

Citation Context

...rocessing queries on U-relations. In our experiments we show that query evaluation on U-relations scales to large amounts of data with high degrees of uncertainty. 1 Introduction Several recent works =-=[10, 9, 8, 2, 14, 4, 6]-=- aim at developing scalable representation systems and query processing techniques for large collections of uncertain data as they arise in data cleaning, Web data management, and scientific databases...

Citation Context

... w is a fresh domain value for ct. � There are U-relations, however, whose ULDB encodings are necessarily exponential in the arity of the logical relation. This is the case of, e.g., or-set relations =-=[13]-=-, attribute-level representations that can be linearly encoded as U-relations but exponentially as ULDBs. Theorem 5.6. U-relational databases are exponentially more succinct than ULDBs. Both ULDBs and...

Citation Context

...rocessing queries on U-relations. In our experiments we show that query evaluation on U-relations scales to large amounts of data with high degrees of uncertainty. 1 Introduction Several recent works =-=[10, 9, 8, 2, 14, 4, 6]-=- aim at developing scalable representation systems and query processing techniques for large collections of uncertain data as they arise in data cleaning, Web data management, and scientific databases...

Citation Context

...ical approximation techniques for confidence computation. 12 Support for new language constructs. Following our recent investigation on uncertainty-aware language constructs beyond relational algebra =-=[5]-=-, we identified common physical operators needed to implement many primitives for the creation and grouping of worlds. It appears that normalizing sets of ws-descriptors in the sense of Section 4 play...

Citation Context

...rocessing queries on U-relations. In our experiments we show that query evaluation on U-relations scales to large amounts of data with high degrees of uncertainty. 1 Introduction Several recent works =-=[10, 9, 8, 2, 14, 4, 6]-=- aim at developing scalable representation systems and query processing techniques for large collections of uncertain data as they arise in data cleaning, Web data management, and scientific databases...