www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.learn - regex problems

reply "seany" <seany uni-bonn.de> writes:
consider this:


import std.conv, std.algorithm;
import core.vararg;
import std.stdio, std.regex;

void main()
{

     string haystack = "ID : generateWorld;
						    Position : { &
										      {ID : \" absolute ; Coordinate : , NULL OMEGA;}
										      {ID : \" inclusion ; Coordinate : UNDEF;}
										      {ID : \" subarc; Coordinate : , NULL OMEGA;	}
								      }; ID : ";
								
     // thus, something like *{B}* can not end here,
     // but something like X can start here.

     string needle = 
"(?<!(([.\n\r])*(\\{)([.\n\r])*))(ID(\\p{White_Space})*:(\\p{White_Space})*)(?!(([.\n\r])*(\\})([.\n\r])*))";

     auto r = regex(needle, "g");
     auto m = matchAll(haystack, r);

     foreach (c; m)
       writeln(c.hit);

}


So let us break up needle:

(
?<!
   (
     ([.\n\r])*(\\{)([.\n\r])*
   )
)

Do not match somthing, that may contain a "*{*" as a leading 
match, * this time means any character, including \n and \r

(ID(\\p{White_Space})*:(\\p{White_Space})*)

however, look for the form : "ID" <few blank spaces> ":" < more 
blank spaces>

(?!(([.\n\r])*(\\})([.\n\r])*))

but no trailing "*}*" as a trailing match.

In haystack, there are two such "ID :" -s. once at the beginning, 
ID : generateWorld. and then the final, last ID

However, this is returning all 5 ID-s as match

what am I doing wrong?
Sep 20 2014
next sibling parent "anonymous" <anonymous example.com> writes:
On Saturday, 20 September 2014 at 15:28:54 UTC, seany wrote:
 In haystack, there are two such "ID :" -s. once at the 
 beginning, ID : generateWorld. and then the final, last ID

 However, this is returning all 5 ID-s as match

 what am I doing wrong?
Prints ID : ID : for me. I'd advise against using regular expressions in this way, though. They are not the proper tool for nested structures. Coming up with correct(!) regexes is probably harder than the alternatives: * using a parser generator like Pegged [1] (haven't used it myself) which supports more powerful grammars than regular expressions, * writing a (recursive descent) parser manually. [1] https://github.com/PhilippeSigaud/Pegged
Sep 20 2014
prev sibling parent "AsmMan" <jckj33 gmail.com> writes:
On Saturday, 20 September 2014 at 15:28:54 UTC, seany wrote:
 consider this:


 import std.conv, std.algorithm;
 import core.vararg;
 import std.stdio, std.regex;

 void main()
 {

     string haystack = "ID : generateWorld;
 						    Position : { &
 										      {ID : \" absolute ; Coordinate : , NULL OMEGA;}
 										      {ID : \" inclusion ; Coordinate : UNDEF;}
 										      {ID : \" subarc; Coordinate : , NULL OMEGA;	}
 								      }; ID : ";
 								
     // thus, something like *{B}* can not end here,
     // but something like X can start here.

     string needle = 
 "(?<!(([.\n\r])*(\\{)([.\n\r])*))(ID(\\p{White_Space})*:(\\p{White_Space})*)(?!(([.\n\r])*(\\})([.\n\r])*))";

     auto r = regex(needle, "g");
     auto m = matchAll(haystack, r);

     foreach (c; m)
       writeln(c.hit);

 }


 So let us break up needle:

 (
 ?<!
   (
     ([.\n\r])*(\\{)([.\n\r])*
   )
 )

 Do not match somthing, that may contain a "*{*" as a leading 
 match, * this time means any character, including \n and \r

 (ID(\\p{White_Space})*:(\\p{White_Space})*)

 however, look for the form : "ID" <few blank spaces> ":" < more 
 blank spaces>

 (?!(([.\n\r])*(\\})([.\n\r])*))

 but no trailing "*}*" as a trailing match.

 In haystack, there are two such "ID :" -s. once at the 
 beginning, ID : generateWorld. and then the final, last ID

 However, this is returning all 5 ID-s as match

 what am I doing wrong?
Is this string a JSON string? if so, why not use a proper JSON parsing library? as other already mentioned, this kind of data isn't good to parse using regex... write small routines to parse that data instead of. It isn't more hard than make it working using regexp. Seriously.
Sep 20 2014