/// <summary>
/// Creates a <seealso cref="TermContext"/> from a top-level <seealso cref="IndexReaderContext"/> and the
/// given <seealso cref="Term"/>. this method will lookup the given term in all context's leaf readers
/// and register each of the readers containing the term in the returned <seealso cref="TermContext"/>
/// using the leaf reader's ordinal.
/// <p>
/// Note: the given context must be a top-level context.
/// </summary>
public static TermContext Build(IndexReaderContext context, Term term)
{
Debug.Assert(context != null && context.IsTopLevel);
string field = term.Field();
BytesRef bytes = term.Bytes();
TermContext perReaderTermState = new TermContext(context);
//if (DEBUG) System.out.println("prts.build term=" + term);
foreach (AtomicReaderContext ctx in context.Leaves)
{
//if (DEBUG) System.out.println(" r=" + leaves[i].reader);
Fields fields = ctx.AtomicReader.Fields;
if (fields != null)
{
Terms terms = fields.Terms(field);
if (terms != null)
{
TermsEnum termsEnum = terms.Iterator(null);
if (termsEnum.SeekExact(bytes))
{
TermState termState = termsEnum.TermState();
//if (DEBUG) System.out.println(" found");
perReaderTermState.Register(termState, ctx.Ord, termsEnum.DocFreq(), termsEnum.TotalTermFreq());
}
}
}
}
return(perReaderTermState);
}
/// <summary>
/// Returns the number of documents containing the term
/// <code>t</code>. this method returns 0 if the term or
/// field does not exists. this method does not take into
/// account deleted documents that have not yet been merged
/// away.
/// </summary>
public override sealed long TotalTermFreq(Term term)
{
Fields fields = Fields;
if (fields == null)
{
return(0);
}
Terms terms = fields.Terms(term.Field());
if (terms == null)
{
return(0);
}
TermsEnum termsEnum = terms.Iterator(null);
if (termsEnum.SeekExact(term.Bytes()))
{
return(termsEnum.TotalTermFreq());
}
else
{
return(0);
}
}
/// <summary>
/// Returns <seealso cref="DocsAndPositionsEnum"/> for the specified
/// term. this will return null if the
/// field or term does not exist or positions weren't indexed. </summary>
/// <seealso cref= TermsEnum#docsAndPositions(Bits, DocsAndPositionsEnum) </seealso>
public DocsAndPositionsEnum TermPositionsEnum(Term term)
{
Debug.Assert(term.Field() != null);
Debug.Assert(term.Bytes() != null);
Fields fields = Fields;
if (fields != null)
{
Terms terms = fields.Terms(term.Field());
if (terms != null)
{
TermsEnum termsEnum = terms.Iterator(null);
if (termsEnum.SeekExact(term.Bytes()))
{
return(termsEnum.DocsAndPositions(LiveDocs, null));
}
}
}
return(null);
}
// DocValues updates
private void ApplyDocValuesUpdates <T1>(IEnumerable <T1> updates, ReadersAndUpdates rld, SegmentReader reader, DocValuesFieldUpdates.Container dvUpdatesContainer) where T1 : DocValuesUpdate
{
lock (this)
{
Fields fields = reader.Fields;
if (fields == null)
{
// this reader has no postings
return;
}
// TODO: we can process the updates per DV field, from last to first so that
// if multiple terms affect same document for the same field, we add an update
// only once (that of the last term). To do that, we can keep a bitset which
// marks which documents have already been updated. So e.g. if term T1
// updates doc 7, and then we process term T2 and it updates doc 7 as well,
// we don't apply the update since we know T1 came last and therefore wins
// the update.
// We can also use that bitset as 'liveDocs' to pass to TermEnum.docs(), so
// that these documents aren't even returned.
string currentField = null;
TermsEnum termsEnum = null;
DocsEnum docs = null;
//System.out.println(Thread.currentThread().getName() + " numericDVUpdate reader=" + reader);
foreach (DocValuesUpdate update in updates)
{
Term term = update.Term;
int limit = update.DocIDUpto;
// TODO: we traverse the terms in update order (not term order) so that we
// apply the updates in the correct order, i.e. if two terms udpate the
// same document, the last one that came in wins, irrespective of the
// terms lexical order.
// we can apply the updates in terms order if we keep an updatesGen (and
// increment it with every update) and attach it to each NumericUpdate. Note
// that we cannot rely only on docIDUpto because an app may send two updates
// which will get same docIDUpto, yet will still need to respect the order
// those updates arrived.
if (!term.Field().Equals(currentField))
{
// if we change the code to process updates in terms order, enable this assert
// assert currentField == null || currentField.compareTo(term.field()) < 0;
currentField = term.Field();
Terms terms = fields.Terms(currentField);
if (terms != null)
{
termsEnum = terms.Iterator(termsEnum);
}
else
{
termsEnum = null;
continue; // no terms in that field
}
}
if (termsEnum == null)
{
continue;
}
// System.out.println(" term=" + term);
if (termsEnum.SeekExact(term.Bytes()))
{
// we don't need term frequencies for this
DocsEnum docsEnum = termsEnum.Docs(rld.LiveDocs, docs, DocsEnum.FLAG_NONE);
//System.out.println("BDS: got docsEnum=" + docsEnum);
DocValuesFieldUpdates dvUpdates = dvUpdatesContainer.GetUpdates(update.Field, update.Type);
if (dvUpdates == null)
{
dvUpdates = dvUpdatesContainer.NewUpdates(update.Field, update.Type, reader.MaxDoc);
}
int doc;
while ((doc = docsEnum.NextDoc()) != DocIdSetIterator.NO_MORE_DOCS)
{
//System.out.println(Thread.currentThread().getName() + " numericDVUpdate term=" + term + " doc=" + docID);
if (doc >= limit)
{
break; // no more docs that can be updated for this term
}
dvUpdates.Add(doc, update.Value);
}
}
}
}
}
private void VerifyTermDocs(Directory dir, Term term, int numDocs)
{
IndexReader reader = DirectoryReader.Open(dir);
DocsEnum docsEnum = TestUtil.Docs(Random(), reader, term.Field(), term.Bytes(), null, null, DocsEnum.FLAG_NONE);
int count = 0;
while (docsEnum.NextDoc() != DocIdSetIterator.NO_MORE_DOCS)
{
count++;
}
Assert.AreEqual(numDocs, count);
reader.Dispose();
}
/// <summary>
/// Returns <seealso cref="DocsAndPositionsEnum"/> for the specified
/// term. this will return null if the
/// field or term does not exist or positions weren't indexed. </summary>
/// <seealso cref= TermsEnum#docsAndPositions(Bits, DocsAndPositionsEnum) </seealso>
public DocsAndPositionsEnum TermPositionsEnum(Term term)
{
Debug.Assert(term.Field() != null);
Debug.Assert(term.Bytes() != null);
Fields fields = Fields;
if (fields != null)
{
Terms terms = fields.Terms(term.Field());
if (terms != null)
{
TermsEnum termsEnum = terms.Iterator(null);
if (termsEnum.SeekExact(term.Bytes()))
{
return termsEnum.DocsAndPositions(LiveDocs, null);
}
}
}
return null;
}
/// <summary>
/// Returns the number of documents containing the term
/// <code>t</code>. this method returns 0 if the term or
/// field does not exists. this method does not take into
/// account deleted documents that have not yet been merged
/// away.
/// </summary>
public override sealed long TotalTermFreq(Term term)
{
Fields fields = Fields;
if (fields == null)
{
return 0;
}
Terms terms = fields.Terms(term.Field());
if (terms == null)
{
return 0;
}
TermsEnum termsEnum = terms.Iterator(null);
if (termsEnum.SeekExact(term.Bytes()))
{
return termsEnum.TotalTermFreq();
}
else
{
return 0;
}
}
/// <summary>
/// Creates a <seealso cref="TermContext"/> from a top-level <seealso cref="IndexReaderContext"/> and the
/// given <seealso cref="Term"/>. this method will lookup the given term in all context's leaf readers
/// and register each of the readers containing the term in the returned <seealso cref="TermContext"/>
/// using the leaf reader's ordinal.
/// <p>
/// Note: the given context must be a top-level context.
/// </summary>
public static TermContext Build(IndexReaderContext context, Term term)
{
Debug.Assert(context != null && context.IsTopLevel);
string field = term.Field();
BytesRef bytes = term.Bytes();
TermContext perReaderTermState = new TermContext(context);
//if (DEBUG) System.out.println("prts.build term=" + term);
foreach (AtomicReaderContext ctx in context.Leaves)
{
//if (DEBUG) System.out.println(" r=" + leaves[i].reader);
Fields fields = ctx.AtomicReader.Fields;
if (fields != null)
{
Terms terms = fields.Terms(field);
if (terms != null)
{
TermsEnum termsEnum = terms.Iterator(null);
if (termsEnum.SeekExact(bytes))
{
TermState termState = termsEnum.TermState();
//if (DEBUG) System.out.println(" found");
perReaderTermState.Register(termState, ctx.Ord, termsEnum.DocFreq(), termsEnum.TotalTermFreq());
}
}
}
}
return perReaderTermState;
}