pos('is',fs) String functiions help

[Peterkl]

Hello guys,

I've been trying to find a specific sub-string within a string, with the "pos" command. It works fine. But there's only one problem that I haven't been able to solve for the moment.  Take a look at my simple code : (the phrase of the txt file is :" this is only a text to see if...")


var
  ft :text ;
  fs :string;
begin
  assign(ft, 'c:\users\peter\desktop\file3.txt');
  reset(ft);
  repeat
   readln(ft,fs);
  until eof (ft);
  writeln(pos('is',fs));
  close(ft);
  readln;
end.         

The output of the program is "3" which is quite normal because the "this" word has the "is" word starting in the 3th position. The question is how can I search for single alone words ? In this case the position of the word would be  5 and not 3.

Cheers

[balazsszekely]

Search for " is " instead of "is".

[marcov]

That will fail on 'Whatever this is, it sucks'  because of the comma. Similarly with other signs at the end of a sentence like ! and .

So you will have to split on words yourself. If you need to support unicode to identify words that is even harder (but not needed for 'is');

[balazsszekely]

@marcov
You're right! My solution is not perfect by far, but to handle all possible scenarios, it would be a tremendous task. Perhaps the easiest way to achieve this is to search for " is" then check if the following char is a " ",  "!",  "." ,  ",", etc..

[fic]

Maybe you can use a StringReplace?
i.e. StringReplace(str,' is','/' ,[rfReplaceAll])?
Mind you : ' is' comes with a space in front, and then continue your search?

[engkin]

A regular expression \bis\b?

Code: Pascal  [Select][+][-]

1. program project1;
2. {$mode objfpc}{$H+}
3. uses
4.   RegExpr;
5.  
6. var
7.   s, u: string;
8.   re: TRegExpr;
9.  
10. begin
11.   re := TRegExpr.Create;
12.   try
13.     re.Expression := '(\bis\b)';
14.     s := 'this is only a text to see if...is,is:isis"is"';
15.     SetLength(u, Length(s));
16.     FillChar(u[1], Length(s), ' ');
17.     if re.Exec(s) then
18.       repeat
19.         u[re.MatchPos[0]] := '-';
20.         WriteLn(re.MatchPos[0]);
21.       until not re.ExecNext;
22.   finally
23.     re.Free;
24.   end;
25.   WriteLn(s);
26.   WriteLn(u);
27.   ReadLn;
28. end.
29.  

[balazsszekely]

@engkin
That's an excellent idea! I think the OP should use this solution.

[Peterkl]

A regular expression \bis\b?

Code: Pascal  [Select][+][-]

1. program project1;
2. {$mode objfpc}{$H+}
3. uses
4.   RegExpr;
5.  
6. var
7.   s, u: string;
8.   re: TRegExpr;
9.  
10. begin
11.   re := TRegExpr.Create;
12.   try
13.     re.Expression := '(\bis\b)';
14.     s := 'this is only a text to see if...is,is:isis"is"';
15.     SetLength(u, Length(s));
16.     FillChar(u[1], Length(s), ' ');
17.     if re.Exec(s) then
18.       repeat
19.         u[re.MatchPos[0]] := '-';
20.         WriteLn(re.MatchPos[0]);
21.       until not re.ExecNext;
22.   finally
23.     re.Free;
24.   end;
25.   WriteLn(s);
26.   WriteLn(u);
27.   ReadLn;
28. end.
29.  

Thank you m8, and thank you for the rest of you guys as well !

[magu]

You may want to try using the following functions from the unit: strutils

ExtractWordPos
WordPosition

Both functions allow you to define delimiters for words (eg ' ', ',') and come with a predefined set of delimters (StdWordDelims)
StdWordDelims = [#0..' ', ',', '.', ';', '/', '\', ':', '''', '"', '`'] + Brackets;

Hope this helps