A database, at its simplest, contains records about objects. Each record in a database corresponds to one object in the collection. For instance, in a card catalog, each card corresponds to a book. Or, in a collection of Christmas ornaments, each record might correspond to a set of ornaments - one record for the box of blue ball ornaments, one record for the box of glass bird ornaments, one record for the box of "12 Days of Christmas" ornaments, etc. (So an "object," in this sense, can be either a literal, single object, or a set of similar things.) The records are set up with a format, or data structure, so that the same kind of information can be included about each object represented in the database. In a card catalog, the format for the cards specifies where the author's name will be typed, where the title will go, etc. In an ornament collection, the structure for the record might specify attributes such as size, color, shape, the number of ornaments in the set, and so on.
The purpose of the database is both to store information about the objects in the collection and to allow you to search for the objects which have particular characteristics (aggregation and discrimination) - all the blue ornaments, for instance, or all the books by A.S. Byatt. The purpose of the data structure is to make the records consistent - they should all contain the same kinds of information about the objects (values), organized in the same way (fields).
In creating a database, the first thing you have to do is decide who will be searching the database and how - what attributes people will want to use to select certain objects and ignore others. You may also want to include some attributes that may not be useful for searching but are useful for other purposes (in a library catalog, the publisher of a book is an example of this kind of attribute). Once you have settled on the attributes to be included, you design the data structure - lay out the order of the fields and specify what kinds of information are to be contained in them. You can also specify what words may be used to express the information (a controlled vocabulary) - whether an ornament's shape is "round" or "spherical," for instance.
After you have completely set up the data structure, you create the actual records for the database -- the representation of each object. Each record will contain a copy of the data structure which you will "fill in" with the appropriate information (values) for an object. Each record corresponds to one object in your collection, and that object may be a single entity or a set of similar entities, as noted above. In other words, the unit of analysis may be a single item or defined sets of items (each individual ornament vs. sets of similar ornaments). Each record begins with an empty copy of the data structure, and you fill in the values for each of the objects in the collection.
Once the records are done, you can then search the database to find the records which meet the particular search criteria. You may search by more than one criterion - yellow color and 2" size, for instance, or yellow color and blue color. You can also indicate alternatives: yellow color or gold color. Boolean Logic, which is part of the database software, allows you to include multiple or alternative criteria in a search.
| Database concept | In DB/TextWorks |
|---|---|
| Design the data structure | Define a New Textbase Edit Fields:
Field Name
Type
Indexing |
| Controlled vocabulary | Validation: Use Validation List
Edit List |
| Records, surrogates | Record: New Record |
| Searching/querying | Search: Query by Example |
| Boolean logic | Boolean Searches |
One Washington Square - San José, California USA, 95192-0029
Phone: 408-924-2490
Contact SLIS
Back