kv

-

kernel handle of type MKernelVector

rts

-

pointer to an RTableSettings structure

rt

-

type ALGEB rtable object

SETTING

TYPE

DEFAULT

EXPLANATION

data_type

int

RTABLE_DAG

type of data in the rtable.

maple_type

ALGEB

anything

type assertion on data

contained in rtables

with data_type=RTABLE_DAG.

subtype

int

RTABLE_ARRAY

type of rtable,

Array, Matrix, or Vector.

storage

int

RTABLE_RECT

storage layout.

p1,p2

M_INT

0

optional parameters used

by some storage layouts.

order

int

RTABLE_FORTRAN

C or Fortran order indexing

(set to RTABLE_FORTRAN,

or RTABLE_C).

read_only

M_BOOL

FALSE

disallow data writes when true.

foreign

M_BOOL

FALSE

true if external program

manages data memory.  If the

pdata parameter to RTableCreate

is non-NULL, this field must be

set to TRUE.

num_dimensions

int

-1

a 5x5 rtable has 2 dimensions.

index_functions

ALGEB

MapleNULL

list of array indexing functions.

attributes

ALGEB

MapleNULL

attributes given to this rtable.

transpose

M_BOOL

FALSE

when true, transpose

the rtable at creation.

fill

ALGEB

0

the default value for

unspecified elements.

SETTING

TYPE

EXPLANATION

RTABLE_DAG

ALGEB

Maple objects

RTABLE_CXDAG

ComplexFloatDAG

complex Maple objects

RTABLE_INTEGER8

char

8-bit integers

RTABLE_INTEGER16

short

16-bit integers

RTABLE_INTEGER32

int

32-bit integers

RTABLE_INTEGER64

long long

64-bit integers

RTABLE_FLOAT32

float

32-bit floating-point numbers

RTABLE_FLOAT64

double

64-bit floating-point numbers

RTABLE_COMPLEX

ComplexFloat64

64-bit complex float pair

SETTING

EXPLANATION

RTABLE_ARRAY

Array

RTABLE_MATRIX

Matrix

RTABLE_COLUMN

column Vector

RTABLE_ROW

row Vector

SETTING

EXPLANATION

RTABLE_SPARSE

sparse format; a table

when data_type is RTABLE_DAG, otherwise

the index and data vector format used

by NAG.  p2 can be set to RTABLE_SPARSE_UPPER

to exclude storing entries below the diagonal

using this format.  Similarly, p2 can be set

to RTABLE_SPARSE_LOWER to exclude entries

above the diagonal.  By default p2 is

RTABLE_SPARSE_DEFAULT, which allows storage

of all entries.

RTABLE_EMPTY

no storage (usually used in

combination with an indexing

function like IdentityMatrix where

the values can be derived without

needing to be stored.

RTABLE_DIAG

only the diagonal entries

are stored.

RTABLE_BAND

banded storage;

where p1 and p2 refer to the

number of non-zero subdiagonals

and superdiagonals.

RTABLE_RECT

dense rectangular storage.

RTABLE_UPPTRI

only the upper triangle

(including the diagonal) is stored.

RTABLE_UPPTRIMINUS

the upper triangle

excluding the diagonal is stored.

RTABLE_UPPTRIPLUS

the upper triangle

plus the one subdiagonal is stored.

RTABLE_LOWTRI

only the lower triangle

(including the diagonal) is stored.

RTABLE_LOWTRIMINUS

the lower triangle

excluding the diagonal is stored.

RTABLE_LOWTRIPLUS

the lower triangle

plus the one superdiagonal is stored.

RTABLE_SCALAR

a single entry is stored.

SETTING

EXPLANATION

RTABLE_ARRAY

Array

RTABLE_MATRIX

Matrix

RTABLE_COLUMN

column Vector

RTABLE_ROW

row Vector

SETTING

EXPLANATION

object of type ALGEB

the name or expression sequence of

indexing functions.

RTABLE_INDEX_ZERO

zero indexing function.

RTABLE_INDEX_BAND

band indexing function.

p1 and p2 must be set to indicate

the number of subdiagonals and

superdiagonals.

RTABLE_INDEX_CONSTANT

constant indexing function.

The fill value is used to denote

the constant stored in every entry

of the rtable.  Usually used in

combination with RTABLE_EMPTY storage.

RTABLE_INDEX_DIAGONAL

diagonal indexing function.

RTABLE_INDEX_SCALAR

scalar indexing function.

Uses the fill value for the main

diagonal entry.  If subtype is a

Vector format, then p1 is additionally

used to indicate which entry is filled

(instead of the diagonal).

RTABLE_INDEX_ANTIHERMITIAN

antihermitian indexing function.

RTABLE_INDEX_HERMITIAN

hermitian indexing function.

RTABLE_INDEX_UPPHESSEN

Hessenberg[upper] indexing function.

RTABLE_INDEX_LOWHESSEN

Hessenberg[lower] indexing function.

RTABLE_INDEX_IDENTITY

identity indexing function.

RTABLE_INDEX_ANTISYMMETRIC

antisymmetric indexing function.

RTABLE_INDEX_SYMMETRIC

symmetric indexing function.

RTABLE_INDEX_LOWTRI

triangular[lower] indexing function.

RTABLE_INDEX_UPPTRI

triangular[upper] indexing function.

RTABLE_INDEX_UNIUPPTRI

triangular[upper,unit] indexing function.

RTABLE_INDEX_UNILOWTRI

triangular[lower,unit] indexing function.

RTABLE_INDEX_UNIT

unit indexing function.

    #include <stdlib.h>

    #include "maplec.h"

    ALGEB M_DECL MyRandomDiagonal( MKernelVector kv, ALGEB *args )

    {

    M_INT argc, n, i;

    RTableSettings rts;

    M_INT bounds[4];

    ALGEB rt;

    INTEGER32 *data, slice;

    FLOAT64 density;

    argc = MapleNumArgs(kv,(ALGEB)args);

    if( argc != 2 ) {

        MapleRaiseError(kv,"two arguments expected");

            return( NULL );

        }

    n = MapleToM_INT(kv,args[1]);

    density = MapleToFloat64(kv,args[2]);

    slice = (INTEGER32)(density * 32767);

    RTableGetDefaults(kv,&rts);

    rts.num_dimensions = 2;

    rts.subtype = RTABLE_MATRIX;

    rts.data_type = RTABLE_INTEGER32;

    rts.storage = RTABLE_DIAG;

    rts.index_functions = (ALGEB)RTABLE_INDEX_DIAGONAL;

    rts.fill = ToMapleInteger(kv,-1);

    rts.order = RTABLE_C;

    rts.read_only = TRUE;

    rts.attributes = ToMapleName(kv,"MyRand",TRUE);

    bounds[0] = 1;

    bounds[1] = n;

    bounds[2] = 1;

    bounds[3] = n;

    rt = RTableCreate(kv,&rts,NULL,bounds);

    data = (INTEGER32*)RTableDataBlock(kv,rt);

    for( i=0; i<n; ++i ) {

        if( rand() % 32767 <= slice )

        data[i] = rand();

    }

    return( rt );

    }