Estimation of the result of a join is possible and can be done using the Cartesian product and intersections of the two relations. It is more complicated than estimating size of a selection.

Size estimation can be done for projection, aggregation, set operations, join operations and also the selection operation.

The size of an aggregation is simply V(A,r) where r is the relation and A is a distinct value since there is one tuple in the aggregation for every distinct value in A.

The union of all records satisfying the individual simple conditions Oi is called as disjunctive selection. It is a form of a complex selection.

The database system catalog stores the number of tuples, number of blocks, size of a tuple in a relation and the blocking factor in it as statistical data.

A random sample must be taken into consideration for statistical analysis because it gives a fair result on the data analysis.

The intersection of all records satisfying the individual simple conditions Oi is called as conjunctive selection. It is a form of a complex selection.

Most of the databases store the distribution of values for each attribute as a histogram.in which the values for the attributes are divided into various ranges.

The analyze command is used to generate statistics on a particular set of relations. The runstats command is also valid but not in SQL. It is used in IBM DB2.

The equi width histogram divides the range of values into equal sized ranges.