Odhad velikosti výsledku vícerozměrného rozsahového dotazu

Abstract

In modern database systems are used multidimensional queries, but also a classic one-dimensional queries. With these multidimensional queries in the database operation to cope with. For efficient processing of these queries, query optimization plan relies on multidimensional selectivity estimation techniques. This technique typically relies on histograms. The basic building block of the histogram is to detect regions of higher density than their surroundings. For the sampling is to obtain data on the population poll respondents in quantitative survey research, statistics and so on. Finding clusters in a multidimensional space may be creating high-quality histograms. We showed principles, concepts and histograms problems and their solutions using different methods. The aim of this thesis is to create a larger solution to estimate the size of the result as one-dimensional selectivity and multidimensional histogram. To tackle the results of a comparison between facts and estimates are applied in the implementation environment programming language.