Headers navigation routines

  One of the main additional complexities caused by the introduction of irregularly sampled data is that navigating the data set, that is computing the pointers into the Data Values File and Header Values File of the records of interest, becomes a non-trivial problem. Since pointers to the data record are stored in the data record number header key of the corresponding header record, the problem is reduced to compute the pointers to the header record, that we call header record number. This task is facilitated if the data set has been gridded. In this case the programmer is provided with two simple headers navigation routines (sep_get_grid_window and sep_put_grid_window), for accessing the header record numbers given the corresponding grid coordinates.

The modality for storing and retrieving the information on the grid coordinates from the database is straightforward. The header record number are accessed for regular sub-windows of the grid coordinates. The sub-windows are specified by using the following windowing parameters for each axis of the Grid Coordinate System: n_wind (length of the window), f_wind (origin of the window) and j_wind (under-sampling rate of the window).