public IEnumerable <IDictionary> Read(Query query)
{
var docIx = new DocIndexReader(DocIndexStream);
var docs = new DocReader(DocStream);
var keyIx = new ValueIndexReader(KeyIndexStream);
var valIx = new ValueIndexReader(ValueIndexStream);
var keyReader = new ValueReader(KeyStream);
var valReader = new ValueReader(ValueStream);
var postingsReader = new PostingsReader(PostingsStream);
var keyHash = query.Term.Key.ToString().ToHash();
var ix = GetIndex(keyHash);
var match = ix.ClosestMatch(query.Term.Value.ToString());
var docIds = postingsReader.Read(match.PostingsOffset, match.PostingsSize);
foreach (var docId in docIds)
{
var docInfo = docIx.Read(docId);
var docMap = docs.Read(docInfo.offset, docInfo.length);
var doc = new Dictionary <IComparable, IComparable>();
for (int i = 0; i < docMap.Count; i++)
{
var kvp = docMap[i];
var kInfo = keyIx.Read(kvp.keyId);
var vInfo = valIx.Read(kvp.valId);
var key = keyReader.Read(kInfo.offset, kInfo.len, kInfo.dataType);
var val = valReader.Read(vInfo.offset, vInfo.len, vInfo.dataType);
doc[key] = val;
}
yield return(doc);
}
}
/// <summary>
/// Find each query term's corresponding index node and postings list and perform "AND", "OR" or "NOT" set operations on them.
/// </summary>
/// <param name="query"></param>
private ScoredResult MapReduce(Query query)
{
Map(query);
var timer = Stopwatch.StartNew();
var result = new PostingsReader(_postingsStream).Reduce(query.ToClauses(), query.Skip, query.Take);
this.Log("map/reduce took {0}", timer.Elapsed);
return(result);
}
