c# - How do I strip valid XML from file that contains excess data and junk -

i have tried parsing regex not xml files being parsed!!

tag names not predefined means, solution should identify start , end of xml tag self

sample data goes this:

<xml data>  random text  <xml data>  random text  

as said in comments hard problem, without knowledge of data. there lot of edge cases.

the code below may work want. attempts find elements ('<element ...' or '<element/>', determine xml content finding end tag. checks trying parse xml , if false rejects xml.

this code written clarity, not performance or structure (it's mainline fall-thru). should give example can use starting point type of parsing described.

using system; using system.linq; using system.text; using system.xml.linq;  namespace messyxml { class program {      const string almostxml = @" @#$%random junk <fruits>   <apples>     pies   </apples>   <pears>     tarts   </pears> </fruits> junk fruits apples , pears  can made pies , tarts.  think pears<apples.  edge case might <parts or /apples> <parts no='123'>   pie plate </parts>";      static void main(string[] args) {         console.writeline("extracting xml from:");         console.writeline(almostxml);         console.writeline();         int = 0;         var validxml = new stringbuilder();         while (i < almostxml.length) {             if (almostxml[i] == '<') { // might xml start                 int ix = almostxml.indexofany(" >\t".toarray(), + 1);                  // check space, > , tab, may want                 // include other whitespace chars                  if (ix < 0) {                     ix = almostxml.indexof("/>", + 1); // might have <element/>                     if (ix >= 0) {                         // check if element name valid                         var xml = almostxml.substring(i, (ix + 2) - i);                         try {                             // see if xml                             var doc = xdocument.parse(xml);                             validxml.appendline(xml);                             = ix + 2;                             continue; // next iteration of while                              }                         catch (system.xml.xmlexception) {                             // nothing                             }                         }                     ix = almostxml.indexof(">", + 1); // might have <element/>                     }                 else { // found <element ...                     var ix2 = almostxml.indexof('>', ix); // '>'                     if (ix2 >= 0) {                         // build end tag                         var endtag = "</" + almostxml.substring(i + 1, (ix - i) - 1) + ">";                         var endix = almostxml.indexof(endtag, ix2);                         if (endix >= 0) {                             var xml = almostxml.substring(i, (endix + endtag.length) - i);                             try {                                 // see if xml                                 var doc = xdocument.parse(xml);                                 validxml.appendline(xml);                                 = (endix + endtag.length);                                 continue; // next iteration of while                                 }                             catch (system.xml.xmlexception) {                                 // nothing                                 }                             }                         }                     }                 }             i++;             }         console.writeline("-----------");         console.writeline("valid xml found:");         console.writeline(validxml.tostring());         console.readkey();         }     } }