How to split non-delimited email-addresses? - delphi

Question

I have a txt file that contains a large quantity of emails and they are delimited with no space. And I want organize these emails line by line.
So, my question is: how get only the part that separates these emails?
Ex: (foo.txt)
brarabelalima#hotmail.comaracaesporteclube#terra.com.br
And I want like this: (new.txt)
brarabelalima#hotmail.com
aracaesporteclube#terra.com.br
This was my last attempt:
{$APPTYPE CONSOLE}
{$R *.res}
uses
SysUtils,
Windows,
Classes;
procedure arrumarEmailsTXT;
var
Linhas, Colunas:TStringList;
i,j:integer;
arq: TextFile;
begin
Linhas := TStringList.Create;
Colunas := TStringList.Create;
AssignFile(arq, 'new.txt');
Rewrite(arq);
try
Linhas.LoadFromFile('foo.txt');
for i := 0 to Pred(Linhas.Count) do
begin
Colunas.text := StringReplace(Linhas.Strings[i],'',#13,[rfReplaceAll]);
for j := 0 to Pred(Colunas.Count) do
begin
Writeln(arq, Colunas.Strings[j]);
end;
end;
finally
CloseFile(arq);
Linhas.Free;
Colunas.Free;
end;
end;
begin
arrumarEmailsTXT;
end.

Answer 1

First run through the text and create a list with the parts between the #-#
Then, take each entry of the new list and cut the text up from the beginning up to the first period. This will give you a list of the domains those emails have.
Then, you need to fill in the last part of the domains and create a new list base don your experience.
For example, for every 'hotmail' you will have 'hotmail.com' and 'hotmail.co.uk', etc.
After that, you can extract the correct position for each line based on the full domain names.
This however can become very heavy job and take much time and resources. Is your file big? You may need threads to run the job effectively.
Also, knowledge of where the emails come may help you cut down the domains.

Answer 2

SOLUTION:
{$APPTYPE CONSOLE}
{$R *.res}
uses
Winapi.Windows,
Classes,
System.SysUtils;
function ExtractEmails(const Input: String; out Emails: TStringList): Boolean;
var
I : Integer;
Buffer : String;
Tokens : TStringList;
begin
Result := False;
Tokens := TStringList.Create;
Emails := TStringList.Create;
Try
Tokens.Delimiter := #32;
Tokens.DelimitedText := Input;
For I := 0 To Tokens.Count - 1 Do
begin
Buffer := Tokens.Strings[I];
If Length(Buffer) < 5 Then Continue;
If (Pos('#',Buffer) > 1) And
(Pos('.',Buffer) > 3) And
(Pos('.',Buffer) < Length(Buffer)) Then
begin
Emails.Add(Buffer);
end;
end;
Result := True;
Finally
Tokens.Free;
end;
end;
procedure Extrai;
var
Linhas,Emails:TStringList;
arq: TextFile;
begin
Linhas := TStringList.Create;
Emails := TStringList.Create;
try
Linhas.LoadFromFile('foo.txt');
AssignFile(arq, 'emails.txt');
Rewrite(arq);
If ExtractEmails(Linhas.Text,Emails) Then
begin
Writeln(arq,Emails.Text);
end;
finally
CloseFile(arq);
Linhas.Free;
Emails.Free;
end;
end;
begin
Extrai;
end.