Description of the illustration dwhsg_3p_001.eps

This image illustrates the use of approximate aggregates in determining the approximate number of distinct products sold each year.

At the top left is a table labeled containing sales data. Three columns containing data and with the following titles are displayed (from left to right): year, month, and prod_id. The fourth column displays ellipses indicating that there are other columns that are not currently displayed. To the right of the Sales data is the SALES_APPROX_MONTH table. An arrow, from the Sales data to SALES_APPROX_MONTH, connects the two tables. The SALES_DATA_MONTH table contains the following columns, from left to right: year, month, approx_month. The data in the approx_month column contains binary values.

At the bottom of the image, are two tables, from left to right, labeled as follows: “Approximate sales data aggregated by year, month” and “Approximate sales data aggregated by year”. An arrow connects the SALES_APPROX_MONTH table to the table labeled “Approximate sales data aggrgated by year, month”, which contains the following columns: year, month, and to_approx_count_distinct(approx_month). Another arrow connects the SALES_APPROX_MONTH table to the table labeled “Approximate sales data aggrgated by year”, which contains the following columns: year and to_approx_count_distinct(approx_month).