1. SQL/OLAP
Sang-Won Lee Let’s e-Wha! URL: Jul. 12th, 2001
SQL/OLAP
ISO/IEC JTC1 SC32

2. Contents Introduction to OLAP and SQL Issues Current OLAP Solutions
SQL/OLAP
Future OLAP Trends
SQL/OLAP
ISO/IEC JTC1 SC32

3. OLAP On-Line Analytical Processing
E.F. Codd coined the term “OLAP”([1])
Multi-dimensional data model
vs. On-Line Transaction Processing
vs. Data warehouse
SQL/OLAP
ISO/IEC JTC1 SC32

4. Multi-dimensional Data Model
Sales(prod-id,store-id,time-id,qty,amt)
Dimension: Product, Store, Time
Hierarchy:
Product -> Category -> Industry
Store->City -> State -> Country
Date -> Month -> Quarter -> Year
SQL/OLAP
ISO/IEC JTC1 SC32

5. Multi-dimensional Data Model(2)
Operations
roll-up/drill-down
slice/dice
pivot
ranking
comparisons
drill-across
etc.
Example
for each state show me top 10 products based on total sales
what is the percentage growth of Jan-99 total sales over total Jan-98?
for each product show me the quantity shipped and sold
SQL/OLAP
ISO/IEC JTC1 SC32

6. OLAP Operations
Many business operations was hard or impossible to express in SQL
multiple aggregations
comparisons(with aggregation)
reporting features
Be prepared for serious performance penalty
Client and middle-ware tools provide the necessary functionality
OLAP server: ROLAP vs. MOLAP
SQL/OLAP
ISO/IEC JTC1 SC32

7. Multiple Aggregations
Create a 2-dimensional spreadsheets that shows sum of sales by maker as well as model of car
Each subtotal requires a separate aggregate query
SELECT color, make, sum(amt)
FROM sales
GROUP BY color, make
union
SELECT color, sum(amt)
GROUP BY color
SELECT make, sum(amt)
GROUP BY make
SELECT sum(amt)
RED
WHITE
BLUE
Chevy
Ford
By Make
By Color
Sum
Cross Tab
SQL/OLAP
ISO/IEC JTC1 SC32

8. Comparisons Examples:
last year’s sales vs. this year’s sales for each product
requires a self-join
VIEW:
create or replace view v_sales as
select prod-id, year, sum(qty) as sale_sum
from sales
group by prod-id, year;
QUERY:
select cur.year cur_year, cur.sale_cur_sales, last.sum last_sales
from v_sales curr, v_sales last
where curr.year=(last.year+1)
SQL/OLAP
ISO/IEC JTC1 SC32

9. The Data CUBE Relational Operator Generalizes Group By and Aggregates
Sum
Aggregate
RED
WHITE
BLUE
By Color
Sum
Group By
(with total)
RED
WHITE
BLUE
Chevy
Ford
By Make
By Color
Sum
Cross Tab
The Data Cube and
The Sub-Space Aggregates
CHEVY
FORD
1990
1991
1992
1993
By Year
By Make
By Make & Year
RED
WHITE
BLUE
By Color & Year
By Make & Color
Sum
By Color
source:[6]
SQL/OLAP
ISO/IEC JTC1 SC32

10. Getting Sub-totals: ROLLUP Operation
SELECT year, brand, SUM(qty) FROM sales GROUP BY ROLLUP (year, brand);
YEAR BRAND SUM(qty)
1996 Ford Honda
1996 Toyota
1997 Ford …
ROLLUP Operator
The ROLLUP operator can be used obtain sub-totals in a query. ROLLUP grouping is an extension to the GROUP BY clause in a query that produces a sub-totals in addition to the “regular” grouped rows. The sub-totals are rows that contain further aggregates whose values are derived by applying the same set functions that were used to obtain the grouped rows.
In the example, the use of the ROLLUP operator produces the following rows in addition to the sum of AMOUNT for each YEAR-BRAND pair:
Sum of QTY for each year
Grand total of QTY for all the years
If there are 2 years and 3 products as shown in the example, a GROUP BY operation without ROLLUP would produce at most 6 rows (m*n), while the same operation using ROLLUP could result in 9 rows (m*(n+1)+1).
SQL/OLAP
ISO/IEC JTC1 SC32

11. Getting Cross-tabs: CUBE Operation
SELECT year, brand, SUM(amount) FROM sales GROUP BY CUBE (year, brand);
YEAR BRAND SUM(AMOUNT)
1996 Ford Toyota Ford
Ford Honda Toyota
A CUBE grouping is an extension to the GROUP BY within a query that produces a result set that contains sub-totals for every possible combinations of the columns or expressions in the GROUP BY clause. The aggregation of those rows are known as n-dimensional cross-tabulation, and produces cubes as discussed earlier.
In the example shown the CUBE operator would produce productwise sub-totals in addition to all the aggregates produced by the ROLLUP operation.
Distinguishing NULLs: GROUPING Function
In the result of a ROLLUP or CUBE operation, a NULL value is used to represents the “ALL” value - that is, in the total for year-1, a NULL is shown against the product column. This presents a problem where a user cannot differentiate between an actual NULL and a null representing “ALL”. A new function GROUPING is used to distinguish between the two types of NULLs.
Grouping is a set function that returns the value 1 if the value in the column in the row is a NULL that represents the set of all values resulting from a ROLLUP or CUBE operation. Where the NULL represents traditional NULL, the function returns the value 0.
SQL/OLAP
ISO/IEC JTC1 SC32

12. Flexible Grouping: GROUPING_SETS Operator
SELECT year, brand, color, SUM(qty) FROM sales GROUP BY GROUPING_SETS ((year, brand), (brand,color),());
YEAR BRAND COLOR SUM(QTY) Ford Honda Toyota Ford Honda Toyota
Ford Blue Ford Red Honda Blue Toyota Red Toyota White
Year, Brand
Brand, Color
The GROUPING_SETS operator can be used to group multiple unrelated groupings without the need to use set operations.
Grand total
SQL/OLAP
ISO/IEC JTC1 SC32

13. LAG Operator SQL> SELECT timekey, sales 2 LAG(sales, 12) OVER
3 (ORDER BY timekey) AS sales_last_year,
4 (sales - sales_last_year) AS sales_change
5 FROM sales;
TIMEKEY SALES SALES_LAST_YEAR SALES_CHANGE
… … … ...
SQL/OLAP
ISO/IEC JTC1 SC32

14. MOVING Average SELECT time-id, avg(sum(qty)) over (order by time-id
RANGE INTERVAL ‘2’ DAY PRECEDING )
as mvg_avg_sales
from sales
group by time_id ;
SQL/OLAP
ISO/IEC JTC1 SC32