The Data Table Class

This is a class of objects conceptually familiar from many statistical systems, for many years. Similar data types are also implicitly and sometimes explicitly defined in relational database management systems.

Originally called a data matrix in most statistical discussions, the key concepts are of a sequence of columns (variables), each column containing some observable or computable quantity on each of a set of rows (observations). There are usually names associated with the columns, and the same sequence of rows apply to all columns.

A table in a relational database captures an essentially similar concept. More generally, the result of a query in SQL corresponds to a similar concept, even though typical SQL setups don't produce such an object explicitly. Some of the details below are taken to emphasize the database view, e.g., the term data table and the notion of an index variable.

Here is a fragment of IDL sufficient to define common sorts of variables and the structure of the data table.

 /* typedef's for the kinds of variables understood */
typedef sequence<double> NumericVariable;

typedef sequence<long> IntegerVariable;

typedef sequence<boolean> BooleanVariable;

typedef sequence<string> StringVariable;

typedef sequence<short> ShortIntegerVariable;

 /* a sequence of variables is the data part of the table */
typedef sequence<any> DataTableData;

 /* the table has data, column names, and optionally
    an index variable for the rows */
interface DataTable {
	attribute DataTableData data;
	attribute StringVariable columns;
	attribute long useForRows;
};

Simple Explicit Variables Dates The Whole Thing Why are we doing this? Start
Last modified: Sun Jul 5 18:10:48 EDT 1998
© Lucent Technologies.