I've worked with a dates table to find "missed" days, and also used CTES to generate sets of dates on the fly using a tally table to increment some number of days from a given starting date. Obviously having a single date column and a single clustered index require less disk space, and the cte approach required next to nothing in disk, although you pay for it in CPU cycles. But your table extends the concept to hours, months, and quarters. Could you give us a real world example or two using other dimensions besides day?