Abstract: Query selectivity estimation is one of the key issues for query processing and optimization. This paper presents a method based on domain distribution density to establish histograms in which the distribution of buckets is uniform or nearly. It utilizes the histograms to estimate query selectivity. Experimental results indicate that this method gets query selectivity with high precision for low-dimensional data and can estimate high-dimensional data.

Key words: selectivity estimation, histogram, n-dimension hyperrectangle, distribution density

摘要： 查询选择率估计是查询处理和优化中的关键之一。提出一种基于区域分布密度的方法，用于构造直方图，使其每个桶具有均匀分布或近似均匀分布，利用直方图估计查询选择率。实验结果表明，该方法对低维数据估计得到的查询选择率精度较高，并能对高维数据进行估计。

关键词: 选择率估计, 直方图, n维超矩形, 分布密度

CLC Number: 

• TP391.72