Class TSorArray (unit mwSArray)

Inherits from

TObject

TSorArray class

Constructors



Functions

function Add(Item: Pointer): Integer;


procedure Clear;


procedure CombSortW(SorCompare: TSortArrayCompare);

Driver for the " Comb " routine.

destructor Destroy;

Create } { TSorArray

function Last: Pointer;


procedure MergeSort(SorCompare: TSortArrayCompare);

Non-recursive Mergesort.

procedure QuickSort(SorCompare: TSortArrayCompare);

Based on a non-recursive QuickSort from the SWAG-Archive.

function Comb(jumpsize0: Integer; SorCompare: TSortArrayCompare): boolean;

Multipication by 0.

function Get(Index: Integer): Pointer;


procedure Merge(SorCompare: TSortArrayCompare);

Unfortunately the " Merge " routine needs additional memory An Algorithm to perform merging in linear time without extra space is described in: B.

procedure Put(Index: Integer; Item: Pointer);


procedure Expand;


procedure SetCapacity(NewCapacity:Integer);


Properties

property Capacity : Integer


property Count : Integer


property Items : Pointer


property SorArray : PDynArray


Events

Variables

fCapacity : Integer;


FCount : Integer;


FLeftArray : TSubArray;


FRightArray : TSubArray;


FSorArray : PDynArray;


SwapArray : PDynArray;


TempArray : PDynArray;



Constructors


Functions


function Add(Item: Pointer): Integer;


procedure Clear;


procedure CombSortW(SorCompare: TSortArrayCompare);

Driver for the " Comb " routine. Based on routines from the SWAG-Archive. Very fast, for a smaller number of items with large keys " Comb " may outperform Quicksort. ( Only a few thousends


destructor Destroy;

Create } { TSorArray


function Last: Pointer;


procedure MergeSort(SorCompare: TSortArrayCompare);

Non-recursive Mergesort. Very fast, if enough memory available. The number of comparisions used is nearly optimal, about 3/4 of QuickSort. If comparision plays a very more important role than exchangement, it outperforms QuickSort in any case. ( Large keys in pointer arrays, for example text with few short lines. ) From all Algoritms with O(N lg N) it's the only stable, meaning it lefts equal keys in the order of input. This may be important in some cases.


procedure QuickSort(SorCompare: TSortArrayCompare);

Based on a non-recursive QuickSort from the SWAG-Archive. ( TV Sorting Unit by Brad Williams )


function Comb(jumpsize0: Integer; SorCompare: TSortArrayCompare): boolean;

Multipication by 0.76 gives a slightly better result than division by 1.3. Because of the FOR loop it runs faster on arrays starting with one


function Get(Index: Integer): Pointer;


procedure Merge(SorCompare: TSortArrayCompare);

Unfortunately the " Merge " routine needs additional memory An Algorithm to perform merging in linear time without extra space is described in: B. Huang and M. Langston, " Practical In-place Merging ", Communications of the ACM 31(1988), 348-352.


procedure Put(Index: Integer; Item: Pointer);


procedure Expand;


procedure SetCapacity(NewCapacity:Integer);


Properties


property Capacity : Integer


property Count : Integer


property Items : Pointer


property SorArray : PDynArray


Events


Variables


fCapacity : Integer;


FCount : Integer;


FLeftArray : TSubArray;


FRightArray : TSubArray;


FSorArray : PDynArray;


SwapArray : PDynArray;


TempArray : PDynArray;