Regular Expressions in Delphi

What are regular expressions?

Regular Expressions are a way to search and replace patters of text. In a way, these patterns are like using the wildcards '?' and '*' when searching for files, but regular expressions are immensely more powerful than that. In the pattern you can specify whether the string being sought should occur at the beginning or end of the line, which characters are allowed, how many times they can be repeated and many more things.

What are they used for?

Basically, they are used for three purposes:

Powerful text search (and replace)
You can use regular expressions for example to search for a specified record in a database in a much more powerful way than using the SQL LIKE operator.
Data validation
With Regular Expressions you can validate user input to check whether it follows a given format that due to its complexity and variability cannot be checked with masks. For example, you can check whether a string starts optionally with a plus or minus sign, then a sequence of no more than seven digits and optionally a point followed by up to two digits.
Data extraction
Not only you can check whether a string follows a certain pattern, but you can extract arbitrary parts of the text. For example, for a phone number like '+1 (123) 555-9999' you can extract the country code ('1'), area code ('123') and the local number ('555-9999').

TRegExpr freeware library

Want to implement regular expressions in your Delphi applications? A good way to do it is using the TRegExpr library, which is freeware and comes with full source code and is documented in many languages. This library was developed by Andrey V. Sorokin porting the C code from the well-known Henry Spencer's V8-routines (a subset of Perl Regular Expressions) to Object Pascal. The last version of the library is 0.942 and you can get it from "AnSo @ Web" (Sorokin's web site) at:

AnSo @ Web
http://anso.da.ru/
http://anso.virtualave.net/
TRegExpr library
http://anso.virtualave.net/regexpr.zip (~284 Kb)
Help files: (42-55 Kb)
- Russian: http://anso.virtualave.net/RegExpRu.zip
- English: http://anso.virtualave.net/RegExpE.zip
- Bulgarian: http://anso.virtualave.net/regexpbg.zip
- German: http://anso.virtualave.net/RegExpG.zip
- French: http://anso.virtualave.net/RegExpF.zip
- Spanish: http://anso.virtualave.net/RegExpS.zip

Examples

The syntax of V8 regular expressions is well documented in the help file and you can also find many sources of documentation in the Internet, so we are not going to enter into that, but just show a couple of examples of the usage of the library:

uses RegExpr;

procedure TForm1.Button1Click(Sender: TObject);
// Validates the email address in Edit1
begin
  // Warning: this code should not be used to perform actual
  // email validation. You should check the RFC specification.
  // This is just a simplification to show the use of ExecRegExpr.
  if not ExecRegExpr('[\w\d\-\.]+@[\w\d\-]+(\.[\w\d\-]+)+',
      Edit1.Text) then begin
    ShowMessage('The email address is not valid');
    Edit1.SetFocus;
  end else
    ShowMessage('The email address is valid');
end;

procedure TForm1.Button2Click(Sender: TObject);
// Extracts email addresses contained in Memo1
var
  RegExpr: TRegExpr;
begin
  // Warning: this code will not extract all valid email addresses.
  // This is just a simplification to show the use of Exec and Match.
  ListBox1.Clear;
  RegExpr := nil;
  try
    RegExpr := TRegExpr.Create;
    if RegExpr <> nil then begin
      RegExpr.Expression := '[^\w\d\-\.]([\w\d\-\.]+@[\w\d\-]+'
                          + '(\.[\w\d\-]+)+)[^\w\d\-\.]';
      if RegExpr.Exec(Memo1.Text) then
        repeat
          ListBox1.Items.Add(RegExpr.Match[1]);
        until not RegExpr.ExecNext;
    end;
  except
  end;
  RegExpr.Free;
end;

You can find the full source code of this article in the archive that accompanies the Pascal Newsletter #21.