1. Structured Query Language (SQL) How to build something useful
Advanced Topics
And
How to build something useful
MIS Winter 2006

2. Goal for this class Advanced SQL Commands
Understand how to use Relational Technologies to solve “M to N” and other advanced problems.
Understand Relational Database Cost/Benefit analysis.
Be able to write SQL queries to solve advanced problems.

3. Advanced SQL
SQL is at the same time simple and yet deceptively complex.
Worse, simple queries can run very slowly and complex one very quickly.
The trick to learning SQL is just to do it.

4. Sub Queries
Anything enclosed in Parentheses ( ) is done first as a separate query.
It returns a result set that you can use in the “main” query.
Commonly used with the IN operator.

5. IN SELECT * FROM Product WHERE V_Code = 12345 OR V_Code = 12346;
SELECT * FROM Product WHERE V_Code IN (12345, 12346);
SELECT * FROM Product WHERE V_Code IN (SELECT V_CODE FROM VENDOR WHERE State = ‘MI’);

6. Sub Query Hints Sub Queries can be hard to use.
Develop the Sub Query first as a separate query. Make sure it works independently first.
Remember a Sub Query returns a ‘virtual’ table.
Keep sub queries to a minimum.

7. SELECT DISTINCT Listing Unique Values
SELECT DISTINCT P_CODE FROM POLine; 52

8. SELECT TOP* SELECT TOP n FROM nnn WHERE x=y Example:
SELECT TOP 5 TotalPop
FROM County
ORDER BY TotalPop DESC;

9. LIKE
Show all of the data for Products whose Description contains CD (i.e. CD, CD-ROM, CD-RW…)
SELECT * FROM Product WHERE Description LIKE ‘CD%’;
% Multicharacter match (* in Access)
_ Single character match (? In Access)

10. Soundex
There is an “algorithm” that lets you search for words that sound alike.
Capitalize all letters in the word and drop all punctuation marks. Pad the word with rightmost blanks as needed during each procedure step.
Retain the first letter of the word.
Change all occurrence of the following letters to '0' (zero):   'A', E', 'I', 'O', 'U', 'H', 'W', 'Y'.
Change letters from the following sets into the digit given:
1 = 'B', 'F', 'P', 'V'
2 = 'C', 'G', 'J', 'K', 'Q', 'S', 'X', 'Z'
3 = 'D','T'
4 = 'L'
5 = 'M','N'
6 = 'R'
Remove all pairs of digits which occur beside each other from the string that resulted after step (4).
Remove all zeros from the string that results from step 5.0 (placed there in step 3)
Pad the string that resulted from step (6) with trailing zeros and return only the first four positions, which will be of the form <uppercase letter> <digit> <digit> <digit>.

11. Converting Columns to Rows
Test
Code
Jan
Feb
Mar
Apr … 1 20 23 24 25 2 33 56 67 55 3 21 22
Suppose you have a table that has Attributes for the various months and you now need to roll to another year.
SELECT Code, Jan FROM Test UNION ALL SELECT Code, Feb FROM Test UNION ALL SELECT Code, Mar FROM Test UNION ALL SELECT Code, Apr FROM Test;

12. Another Example SELECT 'A' as COLX, COLA as COLY from Example
UNION ALL
SELECT 'B' , COLB
SELECT 'C' , COLC
SELECT 'D' , COLD
order by COLX, COLY
Example
COLA COLB COLC COLD
Result
COLX COLY A A B B C C D D

13. How to Build Something Useful

14. The Issue
Businesses, Health Care facilities, Non-profits, the Military and Governmental agencies run on the movement of information. In other words…
Paperwork

15. The Business of IT
Very often you are given a bunch of seemingly unconnected “artifacts” and you need to make something out of them.
These can take the form of Databases, Files, Spreadsheets, Word Documents but most often pieces of paper.

16. Paperwork
Modeling the structure and the flow of paper or now paper like objects is the nature of what IT people do.
Most of those pieces of paper end up as one or more tables in a relational database.

17. Forms Examples The Purchase Order The Invoice Intake logs Test results
Etc…..

18. Forms and Business Rules
Forms also encode many of our business rules
You may need to represent some of the business rules in your database.
Examples
Code sets
Numbering schemes
etc

19. What is so difficult about this?
What are the “Facts” we find here?

20. Facts about many things
Facts about the PO or Invoice itself
Facts about customers or suppliers
Facts about items sold or shipped
Facts about payment
Facts about shipping
Facts about business process
Misc. facts

21. Lets build a model

22. The Simplest Version Phase 0
PurchaseOrder
What is wrong with this approach?

23. PO Table
What are these facts about?

24. PO Model phase 1 1 M Has POInfo Lines
How much redundancy does this add?

25. Results
You need a new Foreign Key in the Lines table that adds a link to the new Primary Key PONumber
This means you are going to duplicate the PONumber many times in the tables.

26. What else can we do? Look for other “redundancies”
What is the Cost/Benefit of adding new tables to address them?1
New FK FieldSize X N =
%ofDuplicates X Total RecordSize X N
New FK FieldSize = %ofDuplicates
Total RecordSize
1. cost of redundancy only

27. PO Model phase 2 1 N
Has
POInfo
Line M 1
Has
Vendor

28. PO Model phase 3 PartNum PartDescr PONumber LineKey PartNum PONumber 1
POInfo
POLine
Product M 1
VendNum
Vendor
VendNum
VendAddr

29. Where do you put?
Totals and subtotals?
Taxes?
Payment details?

30. Data Dictionary Table Attribute DataType PK/FK On Vendor V_Code
Integer PK
V_Name
Char
Product
P_Code
POInfo
PONum FK
POLine
LineKey

31. Now Build the Tables

32. Data Definition Commands
CREATE TABLE POInfo
(PO_Number FCHAR(5) NOT NULL UNIQUE, V_CODE FCHAR(3) NOT NULL, PRIMARY KEY (PO_Number);
FOREIGN KEY (S_CODE) REFERENCES VENDOR ON DELETE RESTRICT ON UPDATE CASCADE)); 12

33. Data Definition Commands
CREATE TABLE POLine (LineKey FCHAR(5) NOT NULL UNIQUE, PONumber VCHAR(35) NOT NULL, PartNum VCHAR(15) NOT NULL, PRIMARY KEY (LineKey));
FOREIGN KEY (PONumber) REFERENCES POInfo ON DELETE RESTRICT ON UPDATE CASCADE));
FOREIGN KEY (PartNum) REFERENCES Product ON DELETE RESTRICT ON UPDATE CASCADE)); 12

34. Data Definition Commands
CREATE TABLE Vendor (V_CODE FCHAR(5) NOT NULL UNIQUE, V_NAME VCHAR(35) NOT NULL, V_CONTACT VCHAR(15) NOT NULL, V_AREACODE FCHAR(3) NOT NULL, V_PHONE FCHAR(3) NOT NULL, V_STATE FCHAR(2) NOT NULL, v_ORDER FCHAR(1) NOT NULL, PRIMARY KEY (V_CODE)); 12

35. Data Definition Commands
CREATE TABLE PRODUCT( P_CODE VCHAR(10) NOT NULL UNIQUE, P_DESCRIPT VCHAR(35) NOT NULL, P_INDATE DATE NOT NULL, P_ONHAND SMALLINT NOT NULL, P_MIN SMALLINT NOT NULL, P_PRICE DECIMAL(8,2) NOT NULL, P_DISCOUNT DECIMAL(4,1) NOT NULL, V_CODE SMALLINT, PRIMARY KEY (P_CODE), FOREIGN KEY (S_CODE) REFERENCES VENDOR ON DELETE RESTRICT ON UPDATE CASCADE); 13

36. Data Definition Commands
SQL Integrity Constraints
Entity Integrity
PRIMARY KEY
NOT NULL and UNIQUE
Referential Integrity
FOREIGN KEY
ON DELETE
ON UPDATE
Remember the business rules! 14

37. How About “Payment Type”?
How do you handle something like this?
What is it a “fact” about.

38. How about a more Complex Example

39. Sub Sections

40. Modeling Issues Keep the “Facts” right.
Look at the Cost/Benefit relationships
Make sure you get rid of all of the M x N relationships.
KISS = Keep It Simple Students.

41. So now that you built it
We need to query it.
We need to manage it.

43. How would we recreate the Purchase Order in SQL?
It depends on what the Model looks like.

44. Phase 0 revisited
Simple SQL based on One table.

45. Listing the Table Contents
SELECT * FROM PurchaseOrder;
SELECT P_CODE, P_DESCRIPT, P_INDATE, P_ONHAND, P_MIN, P-PRICE, P_DISCOUNT, S_CODE FROM PurchaseOrder;
But we said this One Table approach was too costly! 18

46. PO Model phase 3 PartNum PartDescr PONumber LineKey PartNum PONumber 1
POInfo
POLine
Parts M
VendNum
Vendor
VendNum
VendAddr

47. The Key Question
What do you want to know?

48. Bringing it Together
We now need to get information from more than one table.
This is done with the SQL “Join”.
Joins use the Primary and Foreign Keys to link the tables in the SQL Command together.
Remember!!!
WHERE TABLE1.PRIMARY=TABLE2.FOREIGN

49. More Complex Queries
SELECT PRODUCT.P_DESCRIPT, PRODUCT.P_PRICE, VENDOR.V_NAME, VENDOR.V_CONTACT, VENDOR.V_AREACODE, VENDOR.V_PHONE FROM PRODUCT, VENDOR WHERE PRODUCT.V_CODE = VENDOR.V_CODE
AND VENDOR.V_CODE=1234;

50. Another Cost
When you join two tables you incur a cost due to the Join process.
A Join builds an intermediate file that this the combination of all of the entity instances of the one table combined with all of the entity instances of the other table.
These tables can get very large.

51. WHERE CUSTOMER.AGENT_CODE = AGENT.AGENT_CODE
Join
WHERE CUSTOMER.AGENT_CODE = AGENT.AGENT_CODE

52. Product of the two tables

53. Refine the Join

54. More Complex Queries
Anything in a valid SQL statement can be put in the WHERE part of a joined SQL Query.
More than two tables can be joined.
YOU MUST HAVE THE JOIN EQUALITY
This means they can get quite complex
You can use Sub Queries ( ) and “pretty printing” to help keep things straight.

55. Print a PO FROM VENDOR, PRODUCT, PO, POLine
SELECT PO.PONumber, VENDOR.V_NAME, POLine.Line, PRODUCT.P_CODE, PRODUCT.P_DESCRIPT, PRODUCT.P_PRICE
FROM VENDOR, PRODUCT, PO, POLine
WHERE PO.PONumber = POLine.PONumber AND PRODUCT.P_CODE = POLine.P_CODE AND VENDOR.V_CODE = PO.V_CODE
AND PO.PONumber ="1001";

56. Results PONumber V_NAME Line P_CODE P_DESCRIPT P_PRICE 1001
Rubicon Sis. 1
11QER/31
Power painter, 15 psi., 3-nozzle
$109.99 2
WR3/TT3
Steel matting, 4'x8'x1/6", .5" mesh
$119.95

57. Complications
Any “high fidelity” representation of the PO is going to take multiple passes through the SQL engine.
The best way to do this is with a Report Writing tool. (e.g. Crystal Reports)
Even the report tool is SQL based.

58. Sub Query Example
SELECT PRODUCT.P_DESCRIPT, PRODUCT.P_PRICE, VENDOR.V_NAME
FROM PRODUCT, VENDOR
WHERE PRODUCT.V_CODE = VENDOR.V_CODE
AND VENDOR.V_CODE IN (SELECT VENDOR.V_CODE FROM VENDOR WHERE VENDOR.V_CODE > “1234-TK”) ;

59. Data Input
What can we learn about Data Input from looking at the models?

60. Data Input
The more complex the model the more data validation issues you have.
Here again you can use a sub query to look up data in an existing table.
You also need to SQL Enable data input devices.