2 replies
by sfulle
by shanghai sam
On-Line Analytical Processing is technology that allows business users to perform on-line analyses without(any or much)programmer intermediation. This is done by collecting and processing data in such a way that it answers myriads of possible questions without the need to write complex queries for each question. OLAP programmers don't need to know what the exact user questions will be, just the kinds of metrics and categories that may be used in answering the questions. The automatic processingand indexing, plus a sophisticated user interface does the rest.
OLAP often works by indexing "facts" (basic metrics) against many possible hierarchies of "categories" (ways of arranging data).
Example: retail sales data can include facts like volume of sales, sale price, cost price, discount rate and profit and categories like country/region/store/salesman, supplier/product line/model, year/month/week. Categories often come in a hierarchy that "rolls up" -- weekly sales add up to monthly sales; branch sales add up to regional sales, etc...
The factual data is listed in a large fact table, indexed by each category. Aggregate figures like sums and averages for each category are calculated and also indexed. All these aggregates arealso stored in the fact table as separate entries, making for a large fact table with lots of pre-stored calculations that you can look up just by selecting the right indexes.
An OLAP user interface allows users to select just the data they want to see and "drag" the data around into a suitable format. (So, regions across the table columns and product lines down the rows, or vice versa). Usually you can display the data in a range of chart and graphical formats. "Pivot tables" in spreadsheets work in a similar way.
There's a lot of OLAP product around -- you can find it in databases, business applications and reporting tools.
This discussion is closed to new replies.
Trending IT Employment posts
Sort by Trending