Link to home
Start Free TrialLog in
Avatar of blash
blash

asked on

How to make a checksum procedure?

Hi guys,

I need to make a checksum procedure. Can someone help me?

I have a 15 digit number:

Num:=220840001229500

The last 2 digits are zeroes by default.
To get a "control number" I have to perform the following calculation:

I divide Num with 97, thats:

220840001229500:97=2276701043603.09xxxx

Now I have to use the second digit of this result - the 9.

(97+1)-9=89

So 89 is my "control number",
and now at the end, Num: the last 2 zeroes get replaced with the "control number" and Num2: would be

Num2:=2208400012295/89

Greetings, Blash
Avatar of TheNeil
TheNeil

Try this:

Num := 220840001229500;
rTemp := Num / 97;
iDigit := TRUNC(ABS(FRAC(rTemp)) * 100) DIV 10;
iDigit := (97 + 1) - iDigit;
Num2 := (Num DIV 100) * 100 + iDigit;

You will need to declare Num and Num2 as Int64 variables (Longint ain't big enough)

The Neil =:)
Avatar of blash

ASKER

Dear Neil,

What if my Num: is a DBEdit7.Text?
And how do I declare an Int64

Blash

add to the neil code:
var
       num,num2,idigit:int64;
       rtemp:extended;
begin

Num :=  strtoint(DBEdit7.Text);{220840001229500}
rTemp := Num / 97;
iDigit := TRUNC(ABS(FRAC(rTemp)) * 100) DIV 10;
iDigit := (97 + 1) - iDigit;
Num2 := (Num DIV 100) * 100 + iDigit;
end;
Exactly as f15iaf says although personally I'd check the StrToInt conversion:

TRY
  Num := StrToInt(DBEdit7.Text);
EXCEPT
  //Do something useful
END;

The Neil =:)
Avatar of blash

ASKER

The code works,

but I get a Error message:

EConvertError with message "220401000207585" is not a valid integer value.

Why?
Ahhh. IntToStr can't handle the size of the value. What you need to do is either write your own routine to do it or use this one


  FUNCTION BigStrToInt(sValue : STRING): INT64;
  VAR
    iCount : LONGINT;
  BEGIN
    Result := 0;
    FOR iCount := 1 TO Length(sValue)
    DO
      Result := Result * 10 + StrToInt(sValue[iCount]);
  END;

It then just gets called in the same way as IntToStr:

VAR
  iNum : INT64;
begin
  iNum := BigStrToInt('220401000207585');
...

The Neil =:)
Avatar of blash

ASKER

Dear The Neil,

I think that your BigStrToInt function works.

But now I get another Error message after running the code block. I get:
EAccessViolation at address...Read of address FFFFFFFF.

I am appending the code now. Can you see something wrong there?

procedure TForm1.Table2BeforePost(DataSet: TDataSet);
var
  Num, Num2, idigit: int64;
  rtemp: extended;
//Od The Neil
  iNum: int64;
//
begin
//moe,ama samo dopola vrsi rabota
//  if Pos('4',DBEdit7.Text)>0 then
  if Copy(DBEdit7.Text,1,1) = '4' then  //Od Mohhamed
  begin
   ShowMessage('Value must not start with a 4');
   Table2.Cancel;
  end;
//Od Neil
iNum:=BigStrToInt(DBEdit7.Text);
rTemp:=iNum/97;
iDigit:=TRUNC(ABS(FRAC(rTemp))*100)DIV 10;
iDigit:=(97+1)-iDigit;
Num2:=(iNum DIV 100)*100+iDigit;
//Od Neil
  if iDigit<>89 then
  begin
   ShowMessage('the bank account is not correct');
   Table2.Cancel;
  end;
end;
Well it depends on which values you're putting into it that are causing the error (the access voliation itself means nothing). I did spot one thing though (but I don't think it's the problem), try adding a Exit to the first IF statement:

 if Copy(DBEdit7.Text,1,1) = '4' then  //Od Mohhamed
 begin
  ShowMessage('Value must not start with a 4');
  Table2.Cancel;
  Exit;
 end;

Beyond that then it's a case of looking at the data. Can you post a value (or values) that are causing problems?

The Neil =:)
Avatar of blash

ASKER

Dear The Neil,

After cheking out my question, I saw that I must redefine it. I'll do that later during the day. I'll try that "Exit;" anyway. Thanks for helping me.

Greetings, Blash
Avatar of blash

ASKER

Hi The Neil,

The 15 digit number is:

Num:=220840001229500

The last 2 digits are not zeroes by default.
To get the "control number" I have to perform the following calculation:

I  divide Num with 97, but I need the reminder, that's:

220840001229500 mod 97, so the reminder is 9.

So now I need this reminder "9" for the "control number":

(97+1)-"reminder"=(97+1)-9=89

So now my Num2: is:

2208400012295/89

If the "reminder" would be "0" then by default the "control number" would be "89".

I hope I got it right now.

Thanks, Blash
Well using the steps you've mentioned the code below will do what you want. I've included loads of comments so hopefully if there is a problem with the approach then we should be able to spot it:

procedure TForm1.Button1Click(Sender: TObject);

 FUNCTION BigStrToInt(sValue : STRING): INT64;
 VAR
   iCount : LONGINT;
 BEGIN
   Result := 0;
   FOR iCount := 1 TO Length(sValue)
   DO
     Result := Result * 10 + StrToInt(sValue[iCount]);
 END;

VAR
  iValue  : INT64;
  iResult : INT64;
  sValue  : STRING;
begin
  sValue := '220840001229500';
  //Get the stribg into an integer
  iValue  := BigStrToInt(sValue);
  //We're only interested in the final two digits for calculating the checksum
  //BUT we need remember the rest of the value (minus the last two digits) so
  //strip off the last two digitd by dividing by 100
  iResult := iValue DIV 100;
  //Divide the COMPLETE value by 97
  iValue  := iValue MOD 97;
  //Calculate the checksum
  iValue  := (97 + 1) - iValue;

  //Add the two checksum digits to the end of the value
  //iResult is the original number but with the last two digits stripped. If we
  //add on two extra zeroes (by multiplying by 100), we can then add the
  //checksum value straight onto the end to give us a final value
  iResult := iResult * 100 + iValue;

  ShowMessage('Result: ' + IntToStr(iResult));
end;

The Neil =:)
Avatar of blash

ASKER

Dear The Neil :)))

The code works prefectly, only one thing:

How can I say in code at the place where:

sValue:='220840001229500';

take care only for the first 13 digits,leave the 14th and 15th alone (thats taken care later with:

iResult:=iResult*100+iValue;

Thanks Blash ;)))

I'm not sure what you mean. As it stands, the code strips off the last two digits (14th and 15th) works out the checksum and then adds the two new values onto the end when it's finished (so you end up with the same number of digits). It's effectively ignoring the 14th and 15th digits and only basing it's calculation on the first 13

The Neil =:(
Avatar of blash

ASKER

Dear The Neil :)))

So sorry that I'm nagging you with my stupid question. You have more then deserved my pityful 50 points. Have you had a nice weekend? I hope this would be the last that you hear from me - concerning this question.

Your code works perfectly for me if I, for example, have this number:
220401000207682, but only if I enter it as:
220401000207600.

In fact I need, after running your procedure that:
sValue: would be the same as iResult:

I hope I've explained it now.

Sincerely, Blash
ASKER CERTIFIED SOLUTION
Avatar of TheNeil
TheNeil

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Avatar of blash

ASKER

That was perfect.
Thank you The Neil.
Hope we'll meet again on EE.

Blash :)))
Anytime

The Neil =:)