GEN_SORTX(3dxml) — Subroutines

Name

gen_sortx − Sort the elements of an indexed vector

FORMAT

GEN_SORTX ( order, status, type, size, n, x, incx, index )

Arguments

ordercharacter∗1
On entry, order specifies the operation to be performed as follows:

If order = ’A’ or ’a’, x is sorted in ascending sequence

If order = ’D’ or ’d’, x is sorted in descending sequence
On exit, order is unchanged.

statuscharacter∗1
On entry, status specifies the operation to be performed as follows:

If status = ’N’ or ’n’, the index vector is new (empty) on input. Elements of data vector x are initially accessed in sequential order.

If stat = ’O’ or ’o’, the index vector is old (full) on input. Elements of data vector x are initially accessed in the order specified by the index vector.
On exit, status is unchanged.

typecharacter∗1
On entry, type specifies the data type of data vector x. The following types are valid:

If type = ’B’ or ’b’, binary (unsigned integer)

If type = ’C’ or ’c’, character string

If type = ’I’ or ’i’, integer (signed)

If type = ’L’ or ’l’, logical - Fortran .TRUE. or .FALSE.

If type = ’R’ or ’r’, IEEE floating point

If type = ’V’ or ’v’, VAXG floating point
On exit, type is unchanged.

sizeinteger∗4
On entry, size specifies the size, in bytes, of each element of data vector x. Valid combinations of type and size are shown below:

If type = ’B’ or ’b’ - size = { 1,2,4,8 }

If type = ’C’ or ’c’ - size = { 0 < size < 65536 }

If type = ’I’ or ’i’ - size = { 1,2,4,8 }

If type = ’L’ or ’l’ - size = { 1,2,4,8 }

If type = ’R’ or ’r’ - size = { 4,8,16 }

If type = ’V’ or ’v’ - size = { 4,8 }
On exit, size is unchanged.

ninteger∗4
On entry, the length of the x vector.
On exit, n is unchanged.

xdata vector
On entry, a length n vector of data to be sorted. Each element of vector x is of type and size specified.
On exit, x is unchanged.

incxinteger∗4
On entry, incx specifies the distance between elements of vector x. incx must be positive.
On exit, incx is unchanged.

indexinteger∗4
On entry, vector index is ignored if status is ’N’ or ’n’, and used as an index vector for data vector x if status is ’O’ or ’o’.
On exit, vector index is overwritten by a permuted vector of indices that may be used to access data vector x in the sorted order specified by order.

Description

GEN_SORTX is a general purpose, in memory, indexed sort routine. GEN_SORTX accepts the following FORTRAN data types:

INTEGER∗1, INTEGER∗2, INTEGER∗4, INTEGER∗8

LOGICAL∗1, LOGICAL∗2, LOGICAL∗4, LOGICAL∗8

REAL∗4, REAL∗8, REAL∗16

CHARACTER∗(∗)

GEN_SORTX also accepts unsigned "binary" integers of 1, 2, 4, or 8 byte sizes.
An indexed radix algorithm is employed to sort the data. For all data types except REAL∗16 and CHARACTER∗(∗), an N∗16 byte work space is acquired from heap storage. For REAL∗16 data, a work space of N∗24 bytes is taken from heap storage. Heap work space required for CHARACTER data is of size N∗((SIZE-1)/8+3)∗8 bytes.
GEN_SORTX does not return a sorted vector. Instead, it returns an index vector index, which may be used to access data in the collating sequence specified by order.
GEN_SORTX is stable and may be used to perform multi-key record sorts.

Example

REAL DATA( 100 )
INTEGER INDEX( 100 )
N = 100
CALL GEN_SORTX( ’Descend’,’New’,’Real’,SIZEOF(DATA(1)),N,DATA,1,INDEX )
DO I=1,N
PRINT ∗,DATA(I),DATA( INDEX(I) )
ENDDO