For some tasks the user may wish to modify the structure of a table or explicitly manage the location and memory allocation of the data. In these cases, the user will need to make use of the RDBColumn interface in addition to the RDB interface. Examples of more advance table manipulations using the RDBColumn interface are also available.
A comment is any line beginning zero or more spaces followed by a sharp sign (#). Characters following the sharp sign are considered to be the body of the comment. Comments may have additional structure within the body. It is possible to have comment variables, also known as header variables. A comment variable is said to exist if the sharp sign is immediately followed by a colon (:). The first word, non-whitespace and non-equal sign (=) characters, following the colon is the comment variable. The value of the comment variable is seperated from the variable by optional space (' ') and tab ('\t') characters and a mandatory equal sign. The variable value is then any text between the equal sign and the newline. It is common to restrict comments to the initial section of the file
The header consists of two lines. The first contains a tab seperated list of column names. Each column name is seperated by a single tab. The second line contains tab seperated column definitions. Definitions consist of an optional numeric value indicating the width, an S(tring), N(umeric), or M(onth) character denoting the type of the data, an optional < or > character indicating left or right justification of data, and an optional text description.
The data consist of tab seperated values, one per column name.
For a more definitive description of the format see http://hea-www.harvard.edu/MST/simul/software/docs/html/perlrdb/rdb.html .