白皮书

More and more organizations are recognizing the benefits of data warehouses and their ability to support critical business intelligence initiatives. Yet many DW/BI projects fail due to poor collaboration on business requirements, faulty designs, and lack of solid metadata. Many of these challenges can be overcome through effective data modeling practices and data modeling tools.

The dictionary definition of a heuristic is “Computers, Mathematics: pertaining to a trial-and-error method of problem solving used when an algorithmic approach is impractical.” which means this is what you do when you do not know what to do next. The term “normalization” comes from Dr. Codd, who borrowed it from the thencurrent political climate in which we were trying to normalize relations with the Soviet block. The goal is to remove redundancy from a database schema and to have the schema maintain some data integrity without procedural code.

Most of us began programming in procedural languages. But SQL is a declarative language. Chris Date described the difference as “What, not How” in one of his books. SQL has three sub-languages; the DDL or Data Declaration Language, the DML or Data Manipulation Language and the DCL or Data Control Language.

The 2000 book Me Talk Pretty One Day (ISBN 978-0316777728) is a collection of essays by American humorist David Sedaris. He moved to a rural part of Normandy, France, and had no idea how to speak or read French at the time. The humor comes from his frustrated attempts to learn the language and culture. The locals knew him as the American idiot who scares their animals by screaming. The warning on electrified fences was in French and a great source of amusement for the neighbors

Dr. Codd’s original relational algebra had eight basic operations. Since RDBMS is based on set theory the first four are traditional set operations: intersection, set difference, union and product These operations are available in SQL, respectively, as INTERSECT, EXCEPT, UNION, and CROSS JOIN respectively.

SQL is the first programming language to have temporal data types in it. If COBOL had done this, we would never have had the "Y2K Crisis" we survived a few years ago. SQL- 92 added temporal data to the language, acknowledging most of what was already in most SQL product by that time. The problem is that each vendor made a trade-off internally. We will get into SQL code later, but since this is an area where people do not have a good understanding, it is better to start with foundations.