1. LIS512 lecture 2 relational databases Thomas Krichel 2010-02-03

2. relational databases A relational database is a set of tables. There may be relations between the tables. Each table has a number of records. Each record has a number of fields. Let us look at this bottom up.

3. the entity We saw last time that an entity is something we want to gather information about. An entity may be a prime focus of interest, like group 1 entities in FRBR. Or an entity may be only interesting in an entity because they have a relationship to entities that are the prime focus, like FRBR group 2 and 3 entities.

4. entity set and entity In fact the talk in FRBR is a bit loose. When FRBR talks about entities, it really means entity sets. FRBR say that work is an entity. It really means that are workS, and each work is an something of interest. “work” is an conceptual entity set, and each work is an entity.

5. attributes An entity has attributes. These are things we want to know about. Say a person has a name, birthday, height and weight. We can use that data to find out of the person is overweight.

6. records A record is a collection of data elements and their values that pertain to one entity that is being described. – Name  Thomas Krichel – Birthday  1965-06-05 Data element names can also be called “attribute names” or “field name”. These terms are pretty much synonymous.

7. record syntax The way to write out a record (it’s syntax) may vary. Attribute: value syntax – Name: Thomas Krichel – Birthday: 1965-06-05 XML syntax – Thomas Krichel – 1965-06-05

8. tables In a database, records are represented as lines in a table. The first line of the table lists the field names. Here is a sample table with two records |name | birthday | | Thomas Krichel| 1965-06-05| | Karl Marx | 1818-05-13|

9. identity Consider the following table |name | birthday | | Thomas Krichel | 1965-06-05| Is the record a duplicate? Without further information, we can not tell.

10. key To be able to find what records describe the same entities, databases use a construct called the key. The key for each record must have a different value. The name of the key field is not important, but it is important that the field is a key.

11. examples Here is a correct example |key|name | birthday | |2 | Thomas Krichel | 1965-06-05| | 1 | Thomas Krichel | 1965-06-05| Here is an incorrect example |key|name | birthday | |1 | Thomas Krichel | 1965-06-05|

12. identity Identity is simple when it comes to person. But when it comes to other entities we are interested in the FBRB world. Identifiers and the handling of them is one of the most difficult issues of digital librarianship.

13. identifier schemes There are many identifier schemes for group 1 entities – ISBN – ISSN – DOI – URL Some of them may cover different manifestations / expression. i.e. is the ISBN for a print book the same as for the ebook?

14. relations Not all database system are relational. The most widely used ones are relational. That means you can build relationships between tables and enforce them. I am coming to this from an old example.

15. example: Movie database ID| title | director| date M1| Gone with the wind | F. Ford Coppola | 1963 M2| Room with a view| Coppola, F Ford | 1985 M3| High Noon| Woody Allan| 1974 M4 | Star Wars| Steve Spielberg | 1993 M5| Alien| Allen, Woody | 1987 M6| Blowing in the Wind| Spielberg, Steven| 1962 Single table No relations between tables, of course

16. problem with this database All data wrong, but this is just for illustration. Name covered inconsistently. There is no way to find films by Woody Allan without having to go through all spelling variations. Mistakes are difficult to correct. We have to wade through all records, a masochist’s pleasure.

17. Better movie database ID| title |director| year M1| Gone with the wind | D1| 1963 M2| Room with a view| D1| 1985 M3| High Noon| D2| 1974 M4 | Star Wars| D3| 1993 M5| Alien| D2 | 1987 M6| Blowing in the Wind| D3| 1962 ID |director name| birth year D1| Ford Coppola, Francis| 1942 D2| Allan, Woody| 1957 D3| Spielberg, Steven| 1942

18. Relational database We have a one to many relationship between directors and film – Each film has one director – Each director has produced many films Here it becomes possible for the computer – To know which films have been directed by Woody Allen – To find which films have been directed by a director born in 1942

19. enforcing relationships Relational database software has ways to enforce relationships. So when you change the record of M5 to say it was directed by director D9, it will complain that no such director has been defined.

20. Many-to-many relationships Each film has one director, but many actors star in it. Relationship between actors and films is a many to many relationship. Here are a few actors ID| sex| actor name | birth year A1| f| Brigitte Bardot | 1972 A2| m| George Clooney| 1927 A3| f| Marilyn Monroe| 1934

21. Actor/Movie table actor id| movie id A1| M4 A2| M3 A3| M2 A1| M5 A1| M3 A2| M6 A3| M4 … as many lines as required

22. relational databases Relational databases are powerful within organizations that have a relatively centralized command and control structure, i.e. within a company and / or a government department. The relational database model has problems when we are working in a coordinated fashion bet Imagine the web working an a relational database model!

23. FRBR FRBR really brings the way relational databases work to the library world. But here is the problem: how to implement this. Imagine a totally FRBRized, relational database library look like. It would be fine if there is one Big Library Agency (BLA). What would BLA do?

24. BLA job At a starting BLA would first have to register all the works, and say what a work is. That would include maintaining catalogs of composite works. This job is problematic because of – the vague nature of the work – the scale of the job (who will pay)

25. currently Libraries purchase items. They can look up manifestation data from providers such as OCLC. Finding out what expression this represents is more difficult. Linking it to a work is almost impossible. One would need BLA for that.

26. a web decentralized digital library Something that computer scientists dream of http://www.youtube.com/watch?v=OM6XIICm_qo He completely avoids the idea of the meaning of data. In a centralized organization, the meaning can be dictated. In a decentralized world, the meaning as to be agreed upon.

27. outlook In the next few weeks, we are looking at emerging ways to try to find some agreement – character set – record format Then we are looking at library related standards doing similar things.

28. http://openlib.org/home/krichel Thank you for your attention! Please switch off machines b4 leaving!