A new approach: the tree-table representation

The main idea is to build a hierarchy of OATs (see Section 2.3.4 and [7]). This hierarchy is based on the $\psi$-tree structure. Each row of a OAV table is right connected to the other OAV table. Such a connection implies logical AND relation in between.

OAV tables used in tree-table representation are divided into two kinds: attribute tables and action tables. Attribute tables are the attribute part of a classical OAT, Action tables are the action part. There is one logical limitation. While attribute tables may have as many rows as needed (a number of columns depends on the number of attributes), action tables may have only one row, it means that the specified action, or set of actions if there is more then one column, may have only one value set, which preserves the consistency.

Figure 2.3: An example of a tree-table representation

An example of a tree-table representation is given in Fig 2.3. Please note, that a tree-table representation is similar to Relational-Data-Base (RDB) data representation scheme.

The main features of the tree-table representation are:

• simplicity,
• hierarchical, tree-like representation,
• highly efficient way of visualization with high data density,
• power of the decission table representation,
• analogies to the RDB data representation scheme.