Kdtree_Create

Create the actual tree structure by giving an input array of data.

note

Input data is input_data(1:d,1:N), NOT the other way around.

Optional arguments: If 'dim' is specified, then the tree will only search the first 'dim' components of input_data, otherwise, dim is inferred from SIZE(input_data,1).

If sort is true then output results will be sorted by increasing distance. The default is false, as it is faster to not sort.

If rearrange is true then an internal copy of the data, rearranged by terminal node, will be made for cache friendliness. The default is true, as it speeds searches, but building takes longer, and extra memory is used.

Interface

Interface
example
↢ close

FUNCTION Kdtree2_create(input_data, dim, sort, rearrange) RESULT(mr)
  TYPE(Kdtree2_), POINTER :: mr
  INTEGER, INTENT(IN), OPTIONAL :: dim
  LOGICAL, INTENT(IN), OPTIONAL :: sort
  LOGICAL, INTENT(IN), OPTIONAL :: rearrange
  REAL(kdkind), TARGET :: input_data(:, :)
END FUNCTION

PROGRAM main
USE easifemBase
USE Kdtree2_Module
IMPLICIT NONE

TYPE(kdtree2), POINTER :: kd
REAL(DFP), ALLOCATABLE :: input_data(:, :)

! FUNCTION kdtree2_create(input_data, dim, sort, rearrange) RESULT(mr)
CALL Reallocate(input_data, 3, 100)
CALL RANDOM_NUMBER(input_data)

kd => kdtree2_create(input_data, sort=.TRUE., rearrange=.TRUE.)
! CALL kdtree2_destroy(kd)

END PROGRAM main

Interface​

Interface