Fortran Intrinsic procedures Using PACK to select elements meeting a condition
Example
The intrinsic pack function packs an array into a vector, selecting elements based on a given mask. The function has two forms
PACK(array, mask)
PACK(array, mask, vector)
(that is, the vector argument is optional).
In both cases array is an array, and mask of logical type and conformable with array (either a scalar or an array of the same shape).
In the first case the result is rank-1 array of type and type parameters of array with the number of elements being the number of true elements in the mask.
integer, allocatable :: positive_values(:)
integer :: values(5) = [2, -1, 3, -2, 5]
positive_values = PACK(values, values>0)
results in positive_values being the array [2, 3, 5].
With the vector rank-1 argument present the result is now the size of vector (whcih must have at least as many elements as there are true values in mask.
The effect with vector is to return that array with the initial elements of that array overwritten by the masked elements of array. For example
integer, allocatable :: positive_values(:)
integer :: values(5) = [2, -1, 3, -2, 5]
positive_values = PACK(values, values>0, [10,20,30,40,50])
results in positive_values being the array [2,3,5,40,50].
It should be noted that, regardless of the shape of the argument array the result is always a rank-1 array.
In addition to selecting the elements of an array meeting a masking condition it is often useful to determine the indices for which the masking condition is met. This common idiom can be expressed as
integer, allocatable :: indices(:)
integer i
indices = PACK([(i, i=1,5)], [2, -1, 3, -2, 5]>0)
resulting in indices being the array [1,3,5].
