Understanding records and fields

In order to produce a bibliographic database, one must separate the data into records (individual bibliographic entries) and fields (components of a single entry such as author, title, or publication date). The field structure described here is for a flat database file in which there are no links between entries.

Records. Each record represents a single bibliographic item. Some records will include information about other sources, but these will be related in some clear way to each other (a book review citation, for example, will refer to the book it is about, and citation to an article in a journal symposium may refer to all of the other articles in that symposium). Each record needs to have a unique identifying number assigned to it. In the examples below, taken from the Isis database, the identifiers include both the source (Isis) and a unique number within the Isis database.

Fields. Citations for bibliographic resources can be broken up into various components. In order to create a standardized database, those components need to separated. This can be done in two ways: the components can be divided into separate fields and they can be part of a single field that is subdivided by a unique, invariable subfield separator. Most bibliographic databases use both types of fields. Two examples follow:

A. Single-item fields: These fields contain information that will be treated by the database as a single unit. Below is an example of how a journal article citation can be separated into distinct fields in a database. The citation is “Physis 37 (2000): pp. 439-466.” Each element is separated into a different field.

Journal title field: Physis

Journal volume field: 37

Year of publication field: 2000

Pagination field: 439-466

B. Subdivided multi-part repeating fields: Subdivided fields are sometimes necessary for certain types of data. This is the case with the author field, for example, because many works have more than one author. Each subdivided field must have a subfield separator that separates different parts of the field. In the case of the author field example below, commas separate surnames from given names and semicolons separate people:

Author field: Rodríguez de Romo, Ana Cecilia; Irigoyen Coria, Arnulfo; Hernández Sánchez, M. Teresa

Notes: (1) The subfield separator must be invariable and unique.

(2) The subfield separators (in this case, commas and semicolons) cannot be used for any other purpose. Names that include a comma, for example, “John W. Smith, Jr.” have to be entered consistently (three-part names of this sort must always appear in the same order—usually, as a second element in the name, thus: “Smith, Jr., John W.”). And corporate names that include a comma—such as the Ashmolean Museum, Oxford—will need to be designated as such, by using a standard unique coding such as the phrase “CorpName:” preceding the name of all institutions: “CorpName: Ashmolean Museum, Oxford”.

C. Subdivided multi-part non-repeating fields: Sometimes several types of information will be present in a single field. If the information is entered in a clear and consistent way, and subdivided according to a fixed rule, then it can be separated as needed by the bibliographic engine. In the journal entry in (A) above, for example, one could use a semicolon, as a subfield separator to establish the following rule for the data: journal name; volume #; year; pagination. In this case, the field would look as follows:

Journal full citation field: Physis; 37; 2000; 439-466

Notes: (1) The punctuation must be invariable and unique and must be present even if one of the subfields is empty. Example: A citation of the journal “Lychnos (2003): pp. 155-172” which has no volume number would be formatted in the Journal citation field above as follows:

Journal citation field: Lychnos; ; 2003; 155-172 [note the double semicolons]

(2) Extra information can be added in a logical place as needed. Thus, “Biology and Philosophy 15, no. 5 (2000): pp. 759-772” would be formatted as follows:

Journal citation field: Biology and Philosophy; 15, no. 5; 2000; 759-772 [note that the comma does not act as a subfield separator here]

(3) The subfield separator (in this case, a semicolon), cannot be used for any other purpose, so it is important to define a separator that will be unique in all cases. The journal title Report and Transactions; Devonshire Association for the Advancement of Science, Literature, and Art, for example, would have to be repunctuated to remove the semicolon before it could be used in the Journal citation field that used a semicolon separator.