How to list the replacement results of regular expressions

[tomitomy]

Hi, everybody, I want to perform a regular expression substitution operation, and then list all the places that have been replaced. Is there any way to do that?

such as:

Code: Pascal  [Select][+][-]

1. program Project1;
2.  
3. uses
4.   RegExpr;
5.  
6. var
7.   Str: string;
8. begin
9.   Str := ReplaceRegExpr('([0-9]+).([0-9]+)', '123_45_678_9', '$2---$1', True);
10.   Writeln('Str = ', Str);
11.   Writeln('Replaced List ...');
12. end.

Result like this:

Code: [Select]

Str = 45---123_9---678
45---123, start=0, length=8
9---678, start=9, length=7

[Thaddy]

Yes, but that  is not as easy as you think:
- either needs some ( complex) code written by you
- or needs a  diff engine (there are many) that implement that complex code for you....

It is not for the lazy...

[Martin_fr]

Don't use ReplaceRegExpr

Write your own loop that use regex matching, and find each match.
For each found match:
- replace it (you have start and length)
- add it to your list.

Continue the matching in the next loop iteration, from the end of the replacement. (Hint move the already matched and the replaced parts to a new string, and remove them from the source)

[tomitomy]

Yes, but that  is not as easy as you think:
- either needs some ( complex) code written by you
- or needs a  diff engine (there are many) that implement that complex code for you....

It is not for the lazy...

Thank you, Thaddy, I don't know how to use diff engine, I've never used it. I thought out a mathed to implemente my needs, but I don't know if it meets all the circumstances, Can someone help me check it out?

Code: Pascal  [Select][+][-]

1. program Project1;
2.  
3. {$mode objfpc}{$H+}
4.  
5. uses
6.   Classes,
7.   RegExpr;
8.  
9. var
10.   SourceStr, TargetStr: string;
11.   Pattern, ReplaceStr, ReplacedOne, MiddleString: string;
12.   SroucePos, TargetPos: integer;
13.   expr: TRegExpr;
14.   Found: boolean;
15.  
16. begin
17.   SourceStr := 'abc123_45_678_9_876_54_3_2_1def';
18.   Pattern := '([0-9]+).([0-9]+)';
19.   ReplaceStr := '$2---$1';
20.   SroucePos := 1;
21.   TargetPos := 0;
22.   expr := TRegExpr.Create(Pattern);
23.  
24.   Found := expr.Exec(SourceStr);
25.   while Found do
26.   begin
27.     // Replace one string that was found
28.     ReplacedOne := ReplaceRegExpr(Pattern, expr.Match[0], ReplaceStr, true);
29.     // Get the string not matched(middle string)
30.     MiddleString := copy(SourceStr, SroucePos, expr.MatchPos[0] - SroucePos);
31.     // Get the entire replacement result
32.     TargetStr := TargetStr + MiddleString + ReplacedOne;
33.     // Get the starting position of the middle string
34.     SroucePos := expr.MatchPos[0] + expr.MatchLen[0];
35.     // Get the position of the replaced string
36.     TargetPos := TargetPos + Length(MiddleString);
37.     writeln(ReplacedOne, ' start=', TargetPos, ' lengh=', Length(ReplacedOne));
38.     // Get the position of the next replaced string
39.     TargetPos := TargetPos + Length(ReplacedOne);
40.     Found := expr.ExecNext;
41.   end;
42.   expr.Free;
43.   // Get the last unmatched string
44.   MiddleString := copy(SourceStr, SroucePos, length(SourceStr) - SroucePos + 1);
45.   // Get the entire replacement result
46.   TargetStr := TargetStr + MiddleString;
47.   writeln('str = ',TargetStr);
48. end.

Result:

Code: [Select]

45---123 start=3 lengh=8
9---678 start=12 lengh=7
54---876 start=20 lengh=8
2---3 start=29 lengh=5
str = abc45---123_9---678_54---876_2---3_1def

[tomitomy]

Don't use ReplaceRegExpr

Write your own loop that use regex matching, and find each match.
For each found match:
- replace it (you have start and length)
- add it to your list.

Continue the matching in the next loop iteration, from the end of the replacement. (Hint move the already matched and the replaced parts to a new string, and remove them from the source)

I'm sorry! I just saw your reply (I replied very slowly, every sentence I have to use translation software translated into English, and then check it again, including comments in the code). Thank you, Martin_fr, Thank you for your help!

[tomitomy]

I'm sorry, I did not practice before posting. I originally had a similar idea, but I do not know how to replace only one search result and get the replaced content. I think my idea can not be achieved, so I come here Post for help. When I see Thaddy's reply, I want to write out my rough idea first, and then ask "How to replace only one search result and get replaced content", but when I wrote that, I found that I can use "ReplaceRegExpr(Pattern, expr.Match[0], ReplaceStr, true)" to achieve, so I continue to write it, and then come back to reply, but I do not know if this code will run perfectly correctly, because "expr.Match[0]" loses its context, I do not know if "ReplaceRegExpr" will get exactly the right result.

I'm very sorry, Thaddy and Martin_fr, I will think carefully and practice seriously before posting at next time. thank you both for your help!