Out of Memory error on Nested loop C# -

i looked , got have 1 remaining runtime error.
code following:

while ((line = reader.readline()) != null) {     while (reader.peek() != '\r')     {         datalinestream.add(getword(reader));     }     lucenedb.addupdateluceneindex(new mats_doc( datalinestream));     datalinestream.clear(); } 

the code importing data loop not break , crash out due following

"an unhandled exception of type 'system.outofmemoryexception' occurred in mscorlib.dll"

what appropriate break conditions outer while loop ensure read entire file , break @ end. struggle since need advance next line , need skip first row in spreadsheet. appreciated.

*update*

i clear list of strings since creating doc lucene index , has 14 fields in , not want list large.

my code getword

private string getword(textreader inputdata)         {             string word = "";              while (inputdata.peek() >= 0)             {                 word += (char)inputdata.read();                 if ((word.contains(";"))) break;             }          return word;     } 

i'm guessing getword isn't advancing reader - causing datalinestream massive.

or file massive.

out of interest - why throwing away line? - surely contains text you're trying read on inner loop?

update

there must logic hole somewhere in code, , guess it's either adding -1 or 0 chars list, or inner string in getword (by way, using += on strings in way dreadful performance , memory usage). line terminators not \r 'naked' \n.

that said, solve problems doing this:

string line = reader.readline(); string wordtemp; while(line != null) {   string[] words =      line.split(";".tochararray(), stringsplitoptions.removeemptyentries);   foreach(var word in words)   {     wordtemp = word.trim(); //get rid of whitespace     if(wordtemp.length != 0)       datalinestream.add(wordtemp);   }   lucenedb.addupdateluceneindex(new mats_doc( datalinestream));   datalinestream.clear();   line = reader.readline(); } 

as - don't throw away line read - use it, exploiting fact strips newlines out too. use string.split crack out each of words, trimming them in process rid of trailing/leading whitespace might have.