Presentation is loading. Please wait.

Presentation is loading. Please wait.

On-Line Analytic Processing

Similar presentations


Presentation on theme: "On-Line Analytic Processing"— Presentation transcript:

1 On-Line Analytic Processing
Chetan Meshram Class Id:221

2 Agenda Introduction Multidimensional View of OLAP Data Star Schemas
Examples Slicing and Dicing Conclusion References

3 Introduction - OLAP Provides quick answers to analytical queries that are multi-dimensional in nature. Generally involves highly complex queries that use aggregations. OLAP or Decision-support Queries examine large data. Applications: business reporting for sales, marketing, budgeting and forecasting, financial reporting etc.

4 OLAP Applications Common OLAP application uses Warehouse of sales data
Queries that aggregates sales into groups and identify significant groups Example: Schema for Warehouse: Sales(serialNo, date, dealer, price ) Autos(serialNo, model, color) Dealers(name, city, state, phone) A live phonebook When your contacts update their phone numbers you get them right away. Set your preferred number or where you like to be reached. A friend locator See who is in your neighborhood. Tell people you are out of town. Check the local time of your contact before you call. A mobile broadcaster Leave a short message for your friends. Share your mood. And more... Carrier grade application and universal platform. Dial from any country. Integrated with Open Landmark. Save your favorite places. Location, Location, Location See who is in town? Tell people you are out of town by changing your location. Open Landmark "I'm here" utility No GPS? No problem. Your fon11 account is good for Open Landmark and vice versa. Enable the use "I'm here" feature allows you to use Open Landmark to set your current location. Visibility Your location privacy setting. Show your whereabouts on the street level, neighborhood level, or city/metro level. Local Time Check the local time of your contact before you call. (coming soon)

5 OLAP Applications Query: OLTP Query: SELECT state, AVG(price)
FROM Sales, Dealers Where Sales.dealer = Dealers.name AND date>= ‘ ’ Group BY state; Query classifies recent Sales by state of the dealer and touches large amount of data OLTP Query: Bank Deposists, Air Line Reservations Touches only tiny portion of the database Ex: Find price at which auto with serial number 123 was sold, touches only a single tuple of data.

6 Multidimensional OLAP
Fact Table: Multidimensional Space Central relation or collection of data arranged in a multidimensional space or cube Dimensions: car, dealer and date Point represents sale of automobile Dimensions represent properties of sale. Data Cube Date Dealers Cars

7 Multidimensional OLAP
Types: ROLAP: Relational OLAP Data is stored in relations with a specialized structure called ‘Star Schema’. Fact Table contains raw or unaggregated data Other relations contains values along each dimension MOLAP: Multidimensional OLAP A specialized structure called “Data Cube” is used to hold data and its aggregates. Nonrelational operators implemented by system.

8 Star Schemas Schema for the fact table which links to other relations called “dimension tables”. Fact table is at the centre of the “star” whose points are the dimension tables. Fact table consists of dimensions and dependent attributes Ex: Sales(serialNo, date, dealer, price) serialNo, date and dealer are dimensions Price is dependent attribute Google’s Blog Moblast’s Blog Apple’s iPhone Blog

9 Star Schemas Example: Star Schema:
Dimension tables describe values along each dimension Dimension attribute of fact table is a foreign key of corresponding dimension table Suggest possible groupings in an SQL GROUP BY query

10 Star Schemas Example: Dimension Table:
Autos(serialNo, model, color) Dealers(name, city, state, phone) Fact Table: Sales(serialNo, date, dealer, price) serialNo is a foreign key referencing serialNo of Autos Autos.model and Autos.color can be used to group sales in interesting ways. Breakdown of sales by color, or by dealer.

11 Slicing and Dicing Refers to ability to look at the database from different viewpoints Performed along time axis to analyze trends and find patterns. Choice of partition for each dimension “dices” the data cube into smaller cubes GROUP BY and WHERE clause , a query focuses on particular partitions.

12 Slicing and Dicing Example SELECT color, SUM(price)
FROM Sales NATURAL JOIN Autos WHERE model = ‘Sedan’ GROUP BY color; Query dices by color and slices by model User Patterns and trends can be understood.

13 References http://en.wikipedia.org/wiki/Online_analytical_processing

14 Questions?


Download ppt "On-Line Analytic Processing"

Similar presentations


Ads by Google