How do I create/generate unique ID texts?

[JohnLM 14]

I want to create or calculate a unique 8-digit ID (checksum/hash/etc) for variable-sized texts.

 

I am not sure the ID has to be a certain limited length depending on the length of the actual text, but I would like the ID to be as small as possible.  I definitely don't want a GUID. 

 

EXAMPLE: 

Name    comment
------  --------------------
car     neon.
sumvac  is not my favorite any more but i do like the cooler autom
car     neon.

RESULT: 

ID        Name    comment
--------  ------  ------------
AT1UW72Z  car     neon.
0B1UR7PY  sumvac  is not my favorite any more but i do like the cooler autom
AT1UW72Z  car     neon.

I want to add RESULT to a database table.  Duplicates are okay since I can use a query to remove them from a query/report run.  Or, I can remove them completely (at a later time) and just have unique table entries.  But in general, duplicate entries are acceptable. 

 

Is there anything already included in Delphi that supports this?  I will be working in XE7 and/or 12.2 IDE's.

 

Thanks in advanced. 

 

[FPiette 382]

One possibility is to use a CRC32 applied to the text and then convert the resulting UInt32 to hex-ascii representation.

See

CRC32:

   https://docwiki.embarcadero.com/Libraries/Sydney/en/System.ZLib.crc32

IntToHex:

   https://docwiki.embarcadero.com/Libraries/Sydney/en/System.SysUtils.IntToHex

 

[PeterBelow 238]

7 hours ago, JohnLM said:

I want to create or calculate a unique 8-digit ID (checksum/hash/etc) for variable-sized texts.

 

I am not sure the ID has to be a certain limited length depending on the length of the actual text, but I would like the ID to be as small as possible.  I definitely don't want a GUID. 

 

EXAMPLE: 

Name    comment
------  --------------------
car     neon.
sumvac  is not my favorite any more but i do like the cooler autom
car     neon.

RESULT: 

ID        Name    comment
--------  ------  ------------
AT1UW72Z  car     neon.
0B1UR7PY  sumvac  is not my favorite any more but i do like the cooler autom
AT1UW72Z  car     neon.

I want to add RESULT to a database table.  Duplicates are okay since I can use a query to remove them from a query/report run.  Or, I can remove them completely (at a later time) and just have unique table entries.  But in general, duplicate entries are acceptable. 

 

Is there anything already included in Delphi that supports this?  I will be working in XE7 and/or 12.2 IDE's.

 

Thanks in advanced. 

 

Look at the System.Hash unit, this kind of problem is usually the domain of hashing.

[JohnLM 14]

progress update on this endeavour. . .

 

I've been researching all about Hash and then Hash'ing a string. And I have it sort of working, but there are some issues. 

 

The way I understand it, when a hash is created it is a longword, or cardinal or other value, depending on the implementation of the function doing the hash process. 

 

So, take for instance the following: 

 

// link; https://stackoverflow.com/questions/3690608/simple-string-hashing-function
function StrHash(const st:string): cardinal;
var
 i:integer;
begin
  result:=0;
  for i:=1 to length(st) do
   result:=result*$20844 xor byte(st[i]);
end;

This function returns a Cardinal value. 

 

Then, I wrote the following to help me see what is going on and how its working for my needs. I do not need any security. It is not that level. 

 

procedure TForm1.btnGenClick(Sender: TObject);
var
  s: ansistring;
  c: cardinal;
  i: integer;
  hashLen: integer;
begin
  // c := strhash(edit1.Text); m1.Lines.Add(c.ToString()); // works; // debugging purposes 
  c := StrHash(edit1.Text);       // get/calculate the hash and store it in 'c'
  i:=c;                           // convert cardinal into integer value so i can intTostr(i)
  hashLen := length(intTostr(i)); // get the length and store it.  (But I want an even number length for the Hex output. 
                                  // which is next question, "how do I do that?") *** 
  s := inttohex(i,hashLen);       // convert the hash value, I into a string
  m1.Lines.Add(s);                // and display it--this will be the ID, the hash of the string, any length, say up to 255-chars
end;

 

With regard to the final hash ID, I would like to keep it in a specific length so that the output is clean and not croked/jaggid when I display a bunch in a report/output. 

So, I thinking that I need to calculate "even" numbers from the hashLen variable. So, if hashLen is 9 I want to add 1 and make it 10. If hashLen is 8, nothing to do. 

 

This is what I have so far, and although it does seem work, it has some problems/drawbacks. 

 

[JohnLM 14]

PS; To clarify:  The string (edit1.text) can be any length, from 1-char up to 255-char in length. But the hash value, once converted to a Hex string, I want to be as short as possible, not 255-char.  That is for the edit1.text.  I want the hash value to be around 8 to 12 chars if possible. But if not, then I want to at least "pad" with zero's. 

[Anders Melander 1782]

6 hours ago, JohnLM said:

I want to be as short as possible

"as short as possible" is not a usable specification. The shortest possible key is zero characters long. The longer the key, the fewer key collisions.

 

You need to specify the exact length you want your key to be.

 

Also, are you really sure that you want to hex encode the key? With hex encoding you are wasting half the bits by using 8 bits (i.e. an ansichar) to represent a 4 bit value (i.e. a hex digit).

A more efficient encoding would be something like Base32 (5 bits per byte) or Base64 (6 bits per byte). I believe Delphi has implementations of both. Search the source (or wait for someone here to write what they are, as I'm sure they will do).

[FPiette 382]

8 hours ago, JohnLM said:

I would like to keep it in a specific length so that the output is clean and not croked/jaggid when I display a bunch in a report/output. 

Why don't you use what I proposed? You will always get an 8 characters key and make use only Delphi RTL functions. Please comment on this.

I proposed you use CRC32 which is a kind of hash.

   https://docwiki.embarcadero.com/Libraries/Sydney/en/System.ZLib.crc32

IntToHex:

   https://docwiki.embarcadero.com/Libraries/Sydney/en/System.SysUtils.IntToHex

 

[JohnLM 14]

@FPiette I did consider your suggestion, but I could not understand it.  I tried a second time after your last post and I still could not understand it, how to use it, etc.

 

I think that "buf" is throwing me off. I'm not that good at pointers. I thought buf was an array and was to hold some bytes, so I tryined to move the string into it and, well, I just don't have the skills for this. I'm an amature programmer, and this is my hobby, and I'm not some wiz-kid know-it-all. So, I'm giving it up your suggestion per your link. its 5:32am and I work the night shift (truck unloading work) and I'm exhausted. 

 

However, I found another resource for crc32 and tried that, and it seems to be working. But again, I'm exhausted. I'll play around with it much later. I need sleep. 

[FPiette 382]

3 hours ago, JohnLM said:

I did consider your suggestion, but I could not understand it.  I tried a second time after your last post and I still could not understand it, how to use it, etc.

Here an example:

program Crcr32Demo;
{$APPTYPE CONSOLE}
{$R *.res}
uses
  System.SysUtils,
  System.ZLib;
var
    Buf1 : AnsiString;
    Buf2 : String;
    Crc  : UInt32;
begin
    Buf1 := 'Hello World!';
    Crc := System.ZLib.Crc32(0, PByte(Buf1), Length(Buf1) * Sizeof(Buf1[1]));
    WriteLn('AnsiString      ID=', IntToHex(Crc, 8));

    Buf2 := 'Hello World!';
    Crc := System.ZLib.Crc32(0, PByte(Buf2), Length(Buf2) * Sizeof(Buf2[1]));
    WriteLn('UnicodeString   ID=', IntToHex(Crc, 8));

    ReadLn;
end.

 

The output is:

AnsiString      ID=1C291CA3
UnicodeString   ID=E2106423

 

AnsiString and Unicode string doesn't produce the same result because character code are different (8 bit and 16 bit per character). and CRC32 work at the byte level.