Relational tables consist of entries (keyword, attributes) contained in a dataset or a dataset descriptor. Relational tables are also referred to as "associative arrays" or "dictionaries" (Python). Relational table entries are accessed by the keyword consisting of a string (see Appendix B for the maximum keyword length). The attributes of a keyword are the type, the size, and the data. Section 2 describes the storage scheme of relational table entries. Currently, only arrays of type I, J, E, F, C, Z, K data can be defined. There are no a priori limitations to the number of entries or the number of data elements in an entry to be placed in a relational table. However, the table size limits the number of entries or the number of data elements in an entry.

The data set descriptor is identical to a relational table. However, it is associated to the dataset and accessed with functions specific to the dataset descriptor. The dataset descriptor offers a very useful way of describing the nature of data and of assigning attributes to datasets.

Example: B2000 buckling mode datasets contain the displacements corresponding to specific buckling modes. The descriptor of B2000 buckling mode datasets stores all relevant parameters, such as those listed below: