如何比较两个无类型文件内容是否一样？

Hash算法,校验码一样就说明相同

//增加单元
unit untMD5;// -----------------------------------------------------------------------------------------------
INTERFACE
// ----------------------------------------------------------------------------------------------- uses
Windows;type
MD5Count = array[0..1] of DWORD;
MD5State = array[0..3] of DWORD;
MD5Block = array[0..15] of DWORD;
MD5CBits = array[0..7] of byte;
MD5Digest = array[0..15] of byte;
MD5Buffer = array[0..63] of byte;
MD5Context = record
State: MD5State;
Count: MD5Count;
Buffer: MD5Buffer;
end; procedure MD5Init(var Context: MD5Context);
procedure MD5Update(var Context: MD5Context; Input: pChar; Length: longword);
procedure MD5Final(var Context: MD5Context; var Digest: MD5Digest); function MD5String(M: string): MD5Digest; //MD5String为加密字符串
function MD5File(N: string): MD5Digest; //MD5File为加密这个文件
function MD5Print(D: MD5Digest): string; //MD5Print是将加密后的密文转换成字符串function MD5Match(D1, D2: MD5Digest): boolean; //MD5Match是用来比较密文是否一致// -----------------------------------------------------------------------------------------------
IMPLEMENTATION
// ----------------------------------------------------------------------------------------------- var
PADDING: MD5Buffer = (
$80, $00, $00, $00, $00, $00, $00, $00,
$00, $00, $00, $00, $00, $00, $00, $00,
$00, $00, $00, $00, $00, $00, $00, $00,
$00, $00, $00, $00, $00, $00, $00, $00,
$00, $00, $00, $00, $00, $00, $00, $00,
$00, $00, $00, $00, $00, $00, $00, $00,
$00, $00, $00, $00, $00, $00, $00, $00,
$00, $00, $00, $00, $00, $00, $00, $00
); function F(x, y, z: DWORD): DWORD;
begin
Result := (x and y) or ((not x) and z);
end; function G(x, y, z: DWORD): DWORD;
begin
Result := (x and z) or (y and (not z));
end; function H(x, y, z: DWORD): DWORD;
begin
Result := x xor y xor z;
end; function I(x, y, z: DWORD): DWORD;
begin
Result := y xor (x or (not z));
end; procedure rot(var x: DWORD; n: BYTE);
begin
x := (x shl n) or (x shr (32 - n));
end; procedure FF(var a: DWORD; b, c, d, x: DWORD; s: BYTE; ac: DWORD);
begin
inc(a, F(b, c, d) + x + ac);
rot(a, s);
inc(a, b);
end; procedure GG(var a: DWORD; b, c, d, x: DWORD; s: BYTE; ac: DWORD);
begin
inc(a, G(b, c, d) + x + ac);
rot(a, s);
inc(a, b);
end; procedure HH(var a: DWORD; b, c, d, x: DWORD; s: BYTE; ac: DWORD);
begin
inc(a, H(b, c, d) + x + ac);
rot(a, s);
inc(a, b);
end; procedure II(var a: DWORD; b, c, d, x: DWORD; s: BYTE; ac: DWORD);
begin
inc(a, I(b, c, d) + x + ac);
rot(a, s);
inc(a, b);
end; // ----------------------------------------------------------------------------------------------- // Encode Count bytes at Source into (Count / 4) DWORDs at Target
procedure Encode(Source, Target: pointer; Count: longword);
var
S: PByte;
T: PDWORD;
I: longword;
begin
S := Source;
T := Target;
for I := 1 to Count div 4 do begin
T^ := S^;
inc(S);
T^ := T^ or (S^ shl 8);
inc(S);
T^ := T^ or (S^ shl 16);
inc(S);
T^ := T^ or (S^ shl 24);
inc(S);
inc(T);
end;
end; // Decode Count DWORDs at Source into (Count * 4) Bytes at Target
procedure Decode(Source, Target: pointer; Count: longword);
var
S: PDWORD;
T: PByte;
I: longword;
begin
S := Source;
T := Target;
for I := 1 to Count do begin
T^ := S^ and $ff;
inc(T);
T^ := (S^ shr 8) and $ff;
inc(T);
T^ := (S^ shr 16) and $ff;
inc(T);
T^ := (S^ shr 24) and $ff;
inc(T);
inc(S);
end;
end; // Transform State according to first 64 bytes at Buffer
procedure Transform(Buffer: pointer; var State: MD5State);
var
a, b, c, d: DWORD;
Block: MD5Block;
begin
Encode(Buffer, @Block, 64);
a := State[0];
b := State[1];
c := State[2];
d := State[3];
FF (a, b, c, d, Block[ 0], 7, $d76aa478);
FF (d, a, b, c, Block[ 1], 12, $e8c7b756);
FF (c, d, a, b, Block[ 2], 17, $242070db);
FF (b, c, d, a, Block[ 3], 22, $c1bdceee);
FF (a, b, c, d, Block[ 4], 7, $f57c0faf);
FF (d, a, b, c, Block[ 5], 12, $4787c62a);
FF (c, d, a, b, Block[ 6], 17, $a8304613);
FF (b, c, d, a, Block[ 7], 22, $fd469501);
FF (a, b, c, d, Block[ 8], 7, $698098d8);
FF (d, a, b, c, Block[ 9], 12, $8b44f7af);
FF (c, d, a, b, Block[10], 17, $ffff5bb1);
FF (b, c, d, a, Block[11], 22, $895cd7be);
FF (a, b, c, d, Block[12], 7, $6b901122);
FF (d, a, b, c, Block[13], 12, $fd987193);
FF (c, d, a, b, Block[14], 17, $a679438e);
FF (b, c, d, a, Block[15], 22, $49b40821);
GG (a, b, c, d, Block[ 1], 5, $f61e2562);
GG (d, a, b, c, Block[ 6], 9, $c040b340);
GG (c, d, a, b, Block[11], 14, $265e5a51);
GG (b, c, d, a, Block[ 0], 20, $e9b6c7aa);
GG (a, b, c, d, Block[ 5], 5, $d62f105d);
GG (d, a, b, c, Block[10], 9, $2441453);
GG (c, d, a, b, Block[15], 14, $d8a1e681);
GG (b, c, d, a, Block[ 4], 20, $e7d3fbc8);
GG (a, b, c, d, Block[ 9], 5, $21e1cde6);
GG (d, a, b, c, Block[14], 9, $c33707d6);
GG (c, d, a, b, Block[ 3], 14, $f4d50d87);
GG (b, c, d, a, Block[ 8], 20, $455a14ed);
GG (a, b, c, d, Block[13], 5, $a9e3e905);
GG (d, a, b, c, Block[ 2], 9, $fcefa3f8);
GG (c, d, a, b, Block[ 7], 14, $676f02d9);
GG (b, c, d, a, Block[12], 20, $8d2a4c8a);
HH (a, b, c, d, Block[ 5], 4, $fffa3942);
HH (d, a, b, c, Block[ 8], 11, $8771f681);
HH (c, d, a, b, Block[11], 16, $6d9d6122);
HH (b, c, d, a, Block[14], 23, $fde5380c);
HH (a, b, c, d, Block[ 1], 4, $a4beea44);
HH (d, a, b, c, Block[ 4], 11, $4bdecfa9);
HH (c, d, a, b, Block[ 7], 16, $f6bb4b60);
HH (b, c, d, a, Block[10], 23, $bebfbc70);
HH (a, b, c, d, Block[13], 4, $289b7ec6);
HH (d, a, b, c, Block[ 0], 11, $eaa127fa);
HH (c, d, a, b, Block[ 3], 16, $d4ef3085);
HH (b, c, d, a, Block[ 6], 23, $4881d05);
HH (a, b, c, d, Block[ 9], 4, $d9d4d039);
HH (d, a, b, c, Block[12], 11, $e6db99e5);
HH (c, d, a, b, Block[15], 16, $1fa27cf8);
HH (b, c, d, a, Block[ 2], 23, $c4ac5665);
II (a, b, c, d, Block[ 0], 6, $f4292244);
II (d, a, b, c, Block[ 7], 10, $432aff97);
II (c, d, a, b, Block[14], 15, $ab9423a7);
II (b, c, d, a, Block[ 5], 21, $fc93a039);
II (a, b, c, d, Block[12], 6, $655b59c3);
II (d, a, b, c, Block[ 3], 10, $8f0ccc92);
II (c, d, a, b, Block[10], 15, $ffeff47d);
II (b, c, d, a, Block[ 1], 21, $85845dd1);
II (a, b, c, d, Block[ 8], 6, $6fa87e4f);
II (d, a, b, c, Block[15], 10, $fe2ce6e0);
II (c, d, a, b, Block[ 6], 15, $a3014314);
II (b, c, d, a, Block[13], 21, $4e0811a1);
II (a, b, c, d, Block[ 4], 6, $f7537e82);
II (d, a, b, c, Block[11], 10, $bd3af235);
II (c, d, a, b, Block[ 2], 15, $2ad7d2bb);
II (b, c, d, a, Block[ 9], 21, $eb86d391);
inc(State[0], a);
inc(State[1], b);
inc(State[2], c);
inc(State[3], d);
end;

//接上面
// ----------------------------------------------------------------------------------------------- // Initialize given Context
procedure MD5Init(var Context: MD5Context);
begin
with Context do begin
State[0] := $67452301;
State[1] := $efcdab89;
State[2] := $98badcfe;
State[3] := $10325476;
Count[0] := 0;
Count[1] := 0;
ZeroMemory(@Buffer, SizeOf(MD5Buffer));
end;
end; // Update given Context to include Length bytes of Input
procedure MD5Update(var Context: MD5Context; Input: pChar; Length: longword);
var
Index: longword;
PartLen: longword;
I: longword;
begin
with Context do begin
Index := (Count[0] shr 3) and $3f;
inc(Count[0], Length shl 3);
if Count[0] < (Length shl 3) then inc(Count[1]);
inc(Count[1], Length shr 29);
end;
PartLen := 64 - Index;
if Length >= PartLen then begin
CopyMemory(@Context.Buffer[Index], Input, PartLen);
Transform(@Context.Buffer, Context.State);
I := PartLen;
while I + 63 < Length do begin
Transform(@Input[I], Context.State);
inc(I, 64);
end;
Index := 0;
end else I := 0;
CopyMemory(@Context.Buffer[Index], @Input[I], Length - I);
end; // Finalize given Context, create Digest and zeroize Context
procedure MD5Final(var Context: MD5Context; var Digest: MD5Digest);
var
Bits: MD5CBits;
Index: longword;
PadLen: longword;
begin
Decode(@Context.Count, @Bits, 2);
Index := (Context.Count[0] shr 3) and $3f;
if Index < 56 then PadLen := 56 - Index else PadLen := 120 - Index;
MD5Update(Context, @PADDING, PadLen);
MD5Update(Context, @Bits, 8);
Decode(@Context.State, @Digest, 4);
ZeroMemory(@Context, SizeOf(MD5Context));
end; // ----------------------------------------------------------------------------------------------- // Create digest of given Message
function MD5String(M: string): MD5Digest;
var
Context: MD5Context;
begin
MD5Init(Context);
MD5Update(Context, pChar(M), length(M));
MD5Final(Context, Result);
end; // Create digest of file with given Name
function MD5File(N: string): MD5Digest;
var
FileHandle: THandle;
MapHandle: THandle;
ViewPointer: pointer;
Context: MD5Context;
begin
MD5Init(Context);
FileHandle := CreateFile(pChar(N), GENERIC_READ, FILE_SHARE_READ or FILE_SHARE_WRITE,
nil, OPEN_EXISTING, FILE_ATTRIBUTE_NORMAL or FILE_FLAG_SEQUENTIAL_SCAN, 0);
if FileHandle <> INVALID_HANDLE_VALUE then try
MapHandle := CreateFileMapping(FileHandle, nil, PAGE_READONLY, 0, 0, nil);
if MapHandle <> 0 then try
ViewPointer := MapViewOfFile(MapHandle, FILE_MAP_READ, 0, 0, 0);
if ViewPointer <> nil then try
MD5Update(Context, ViewPointer, GetFileSize(FileHandle, nil));
finally
UnmapViewOfFile(ViewPointer);
end;
finally
CloseHandle(MapHandle);
end;
finally
CloseHandle(FileHandle);
end;
MD5Final(Context, Result);
end; // Create hex representation of given Digest
function MD5Print(D: MD5Digest): string;
var
I: byte;
const
Digits: array[0..15] of char =
('0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'a', 'b', 'c', 'd', 'e', 'f');
begin
Result := '';
for I := 0 to 15 do Result := Result + Digits[(D[I] shr 4) and $0f] + Digits[D[I] and $0f];
end; // ----------------------------------------------------------------------------------------------- // Compare two Digests
function MD5Match(D1, D2: MD5Digest): boolean;
var
I: byte;
begin
I := 0;
Result := TRUE;
while Result and (I < 16) do begin
Result := D1[I] = D2[I];
inc(I);
end;
end; end. //以下是调用的部分
uses
  untMD5;var
  b: Boolean;
begin
  b := MD5Match(MD5File('c:\1.xml'),MD5File('c:\2.xml'));
  if b then
    ShowMessage('相同')
  else
    Showmessage('不同');
end;

没那么复杂。
有几种方式可以达到，WIN32 API、Pascal、流，基本步骤都是：
1、打开文件；
2、比较文件长度，如果不同则不用比较内容了；
3、如果文件长度相同，则读入内容进行比较；
4、关闭文件。下面介绍WIN32 API的方式：
1、打开文件：OpenFile (FileOpen)
2、取得文件长度进行比较：GetFileSizeEx
3、读入文件内容进行比较：ReadFile (FileRead), CompareMem；应该分批分次读入；
4、关闭文件：CloseHandle (FileClose)

支持：agui(阿贵: 高级图形用户界面)
个人认为用流进行比较最好。

帮你写一个吧，放两个SaveDialog，分别为SaveDialog1和SaveDialog2，然后放一个按钮Button1，点击按钮Button1后会出现打开文件的对话框，连续装入两个要比较的文件就可以看到结果了。
procedure TForm1.Button1Click(Sender: TObject);
var
  file1, file2: string;
function CompFile(f1, f2: string): boolean;
var
  fs1, fs2: TFileStream;
  fSize1, fSize2, i: integer;
  b1, b2: byte;
begin
  fs1 := TFileStream.Create(f1,fmOpenRead);
  fs2 := TFileStream.Create(f2,fmOpenRead);
  fSize1 := fs1.Size;
  fSize2 := fs2.Size;
  Result := false;
  if fSize1<>fSize2 then exit;
  for i:=0 to fSize1 do
  begin
    fs1.Read(b1,1);
    fs2.Read(b2,1);
    if b1<>b2 then exit;
  end;
  Result := true;
end;
begin
  if Not(OpenDialog1.Execute) then exit;
  if Not(OpenDialog2.Execute) then exit;
  file1 := OpenDialog1.FileName;
  file2 := OpenDialog2.FileName;
  if CompFile(file1, file2) then
    ShowMessage('两个文件相同')
  else
    ShowMessage('两个文件不同');
end;

修改一下，刚才两个TFileStream忘了Free了procedure TForm1.Button1Click(Sender: TObject);
var
  file1, file2: string;
function CompFile(f1, f2: string): boolean;
var
  fs1, fs2: TFileStream;
  fSize1, fSize2, i: integer;
  b1, b2: byte;
begin
  fs1 := TFileStream.Create(f1,fmOpenRead);
  fs2 := TFileStream.Create(f2,fmOpenRead);
  fSize1 := fs1.Size;
  fSize2 := fs2.Size;
  Result := false;
  if fSize1=fSize2 then
  begin
    for i:=0 to fSize1 do
    begin
      fs1.Read(b1,1);
      fs2.Read(b2,1);
      if b1<>b2 then break;
    end;
    if (i=fSize1+1) then Result := true;
  end;
  fs1.Free;
  fs2.Free;
end;
begin
  if Not(OpenDialog1.Execute) then exit;
  if Not(OpenDialog2.Execute) then exit;
  file1 := OpenDialog1.FileName;
  file2 := OpenDialog2.FileName;
  if CompFile(file1, file2) then
    ShowMessage('两个文件相同')
  else
    ShowMessage('两个文件不同');
end;

to cuteant:
一个字节一个字节比较？
虽然我也是菜鸟，但我还是不敢尝试用这种效率最低的方法比较两个100M或以上的文件。

to liyinwei(月夜彩虹):
随便写了一点给楼主参考而已，看来楼主应该不是需要比较很大的文件。
就算很大的文件，一个字节一个字节比较也可以啊，只不过不是从文件中读一个比较一个而已，像firstrose(kots)说的那样在缓冲区比较就没问题了。

只在放在缓冲区中不用一个节比较,可以用下面的函数,汇编写的效率应该不错。
Performs a binary comparison of two memory images.UnitSysUtilsCategorycomparison routinesDelphi syntax:function CompareMem(P1, P2: Pointer; Length: Integer): Boolean; assembler;C++ syntax:extern PACKAGE bool __fastcall CompareMem(void * P1, void * P2, int Length);DescriptionCompareMem performs a binary compare of Length bytes of memory referenced by P1 to that of P2. CompareMem returns true if the memory referenced by P1 is identical to that of P2.

to cuteant:
请贴出你认为最快、最好的代码。
我真想知道用流的方法最快能达到什么程度。

楼主：最好你自己写，这样你获得的经验对你很有好处。你写出来后如果有问题，可以贴出来，也可以给我email: [email protected] 。另，
1、一个字节一个字节比效率绝对不是最高的，因为 Intel 的CPU指令中就有按块比的（ REP CMPSx, x=B|W|D ），CompareMem 是现成的利用了这些指令的函数，有现成的何苦不用？
2、流的好处是统一和方便，但是它还是调用了Pascal函数或API（WindowsAPI不可能直接提供Delphi流），因此效率要打些折扣。
顺便贴出部分 TFileStream 的实现：
TFileStream 继承自 THandleStream ，使用的是 THandleStream.Read ；
THandleStream.Read 调用了 FileRead 函数；
FileRead 函数调用了 Windows API: ReadFile 。

调试易

如何比较两个无类型文件内容是否一样？

解决方案 »