Cartesian and Structured Tables
Whenever possible, organize tables to be "Cartesian" or at least "Structured." These tables are much faster to search than "Unstructured" tables.
In a Cartesian table, the entries for the independent variable are equally spaced. In a Structured table, the spacing is predictable but not uniform. The organization of a table is reported in the read-only Table Format property for Table(T), Table(T,P), and Table(Reaction Time) tables.
When STAR-CCM+ reads a table, it first sorts the data points in each dimension and then computes the spacing. For example, in a 2D table let there be Nx points in the first dimension and Ny points in the second dimension. If the total number of specified points is Nx*Ny and if the spacing is uniform in each dimension then the table is tagged Cartesian. If the spacing is non-uniform then it is tagged Structured. If the total number of points does not match Nx*Ny, then it is tagged Unstructured. (Note that even if the spacing is uniform, but the total number of points does not match Nx*Ny, the table is tagged Unstructured, since this situation means that some points are missing in the Nx*Ny grid.)
Examples:
- 1D Cartesian
-
This table is a 1D Cartesian table since the sample points are equally spaced. 1D tables can only be Cartesian or Structured; they cannot be Unstructured.
- 2D Cartesian
-
This table is a 2D Cartesian table since the points are equally spaced along the independent variable.
- 2D Structured
-
This table is a 2D Structured table since the points lie on a Structured grid, although the spacing is not uniform.
- 2D Unstructured
-
This table is a 2D Unstructured table since the points are irregularly spaced along the independent variable.
These categories also apply to higher-dimensional tables. If the independent points lie on an N-dimensional Structured grid, then it is a Structured table or (if the spacing is uniform) a Cartesian table. If points are irregularly spaced in the N-dimensional space, then it is an Unstructured table.