How can I speed up reading the Lucene search results

[rhino1130]

Reading lucene search hits results to xml slow slow. I need the xml
and can't find a way to convert the hits object directly to xml. So
now what?

Since there is no direct way to read it to xml, I have to first
convert the hits to a dataset then to xml. Here is what I am doing...

// execute the search
Hits _oWidgetHits =
_oLuceneIndexSearch.ExecuteSearch(_booleanqqueryWidgetsQuery,
INDEXPATH);

// read hit results
if (_oWidgetHits.Length() > 0)
{
for (int iResultsCount = 0; iResultsCount < _oWidgetHits.Length();
iResultsCount++)
{
// get the document from index
Document _dResultsDoc = _oWidgetHits.Doc(iResultsCount);

// create a new row with the result data
DataRow _rWidgets = _dtSearchResults.NewRow();
_rWidgets["WidgetID"] = _dResultsDoc.Get("WidgetID");
_rWidgets["WidgetTitle"] = _dResultsDoc.Get("WidgetTitle");

// add row
_dtSearchResults.Rows.Add(_rRecipe);
}
}

return _dtSearchResults;

I have tried using the hitIterator but that give me no gains. I am
researching the hitCollector but that seems to be the same concept.

The problem with the speed really is the number of records. I have
reduced the results but that really is only a band-aid.