DSXchange

what is the difference between drill acorss,drill down, drill up?

This is a 101 analytical processing question. Just like your previous post "What is metadata", it seems to be a question one would find on a technical job interview.

I'd suggest you read Ralph Kimballs DataWarehouse Lifecycle Toolkit. It will guide you in the 1000 other questions that will follow.

In summary:

Drilling down is descending grains of aggregate data in search of more detail about the underlying transactions. Drilling up is simply moving in the opposite direction. For example, a base fact table contains order transactions, and aggregate table contains summations of orders grouped by order date and product. If someone was looking at order totals for a given day, their query would rollup the aggregate, losing the product column in the summation. Now, if they spotted a day of interest, they may wish to descend into the underlying grouping of products for that day. Furthermore, they could descend even further into the transactions for a product for that day. This is basic drill-down/drill-up activities.

Drilling across is simply moving around different aggregate groups looking at things of interest. As in the previous example, you're looking at activity for a given product, simply side-stepping to another product is drilling across.