A Understanding Cardinality Estimation using Entropy Maximization

Cardinality estimation is the process of estimating the number of tuples returned by a query. In relational database query optimization, cardinality estimates are key statistics used by the optimizer to choose an (expected) lowest cost plan. As a result of the importance of the problem, there are many sources of statistical information available to the optimizer, e.g., query feedback records [Stillger et al. 2001; Chaudhuri et al. 2008] and distinct value counts [Alon et al. 1996], and many models to capture some portion of the available statistical information, e.g., histograms [Poosala and Ioannidis 1997; Kaushik and Suciu 2009], samples [Haas et al. 1996], and sketches [Alon et al. 1999; Rusu and Dobra 2008]; but on any given cardinality estimation task, each method may return a different (and so, conflicting) estimate. Consider the following cardinality estimation task: