private void MergeTerms()
{
try
{
freqOutput = directory.CreateOutput(segment + ".frq");
proxOutput = directory.CreateOutput(segment + ".prx");
termInfosWriter = new TermInfosWriter(directory, segment, fieldInfos, termIndexInterval);
skipInterval = termInfosWriter.skipInterval;
maxSkipLevels = termInfosWriter.maxSkipLevels;
skipListWriter = new DefaultSkipListWriter(skipInterval, maxSkipLevels, mergedDocs, freqOutput, proxOutput);
queue = new SegmentMergeQueue(readers.Count);
MergeTermInfos();
}
finally
{
if (freqOutput != null)
{
freqOutput.Close();
}
if (proxOutput != null)
{
proxOutput.Close();
}
if (termInfosWriter != null)
{
termInfosWriter.Close();
}
if (queue != null)
{
queue.Close();
}
}
}
internal FormatPostingsTermsWriter(SegmentWriteState state, FormatPostingsFieldsWriter parent)
: base()
{
this.parent = parent;
termsOut = parent.termsOut;
docsWriter = new FormatPostingsDocsWriter(state, this);
}
public FormatPostingsFieldsWriter(SegmentWriteState state, FieldInfos fieldInfos) : base()
{
dir = state.directory;
segment = state.segmentName;
totalNumDocs = state.numDocs;
this.fieldInfos = fieldInfos;
termsOut = new TermInfosWriter(dir, segment, fieldInfos, state.termIndexInterval);
// TODO: this is a nasty abstraction violation (that we
// peek down to find freqOut/proxOut) -- we need a
// better abstraction here whereby these child consumers
// can provide skip data or not
skipListWriter = new DefaultSkipListWriter(termsOut.skipInterval, termsOut.maxSkipLevels, totalNumDocs, null, null);
state.flushedFiles.Add(state.SegmentFileName(IndexFileNames.TERMS_EXTENSION));
state.flushedFiles.Add(state.SegmentFileName(IndexFileNames.TERMS_INDEX_EXTENSION));
termsWriter = new FormatPostingsTermsWriter(state, this);
}
public FormatPostingsFieldsWriter(SegmentWriteState state, FieldInfos fieldInfos):base()
{
dir = state.directory;
segment = state.segmentName;
totalNumDocs = state.numDocs;
this.fieldInfos = fieldInfos;
termsOut = new TermInfosWriter(dir, segment, fieldInfos, state.termIndexInterval);
// TODO: this is a nasty abstraction violation (that we
// peek down to find freqOut/proxOut) -- we need a
// better abstraction here whereby these child consumers
// can provide skip data or not
skipListWriter = new DefaultSkipListWriter(termsOut.skipInterval, termsOut.maxSkipLevels, totalNumDocs, null, null);
SupportClass.CollectionsHelper.AddIfNotContains(state.flushedFiles, state.SegmentFileName(IndexFileNames.TERMS_EXTENSION));
SupportClass.CollectionsHelper.AddIfNotContains(state.flushedFiles, state.SegmentFileName(IndexFileNames.TERMS_INDEX_EXTENSION));
termsWriter = new FormatPostingsTermsWriter(state, this);
}
private void WritePostings(Posting[] postings, System.String segment)
{
IndexOutput freq = null, prox = null;
TermInfosWriter tis = null;
TermVectorsWriter termVectorWriter = null;
try
{
//open files for inverse index storage
freq = directory.CreateOutput(segment + ".frq");
prox = directory.CreateOutput(segment + ".prx");
tis = new TermInfosWriter(directory, segment, fieldInfos, termIndexInterval);
TermInfo ti = new TermInfo();
System.String currentField = null;
for (int i = 0; i < postings.Length; i++)
{
Posting posting = postings[i];
// add an entry to the dictionary with pointers to prox and freq files
ti.Set(1, freq.GetFilePointer(), prox.GetFilePointer(), - 1);
tis.Add(posting.term, ti);
// add an entry to the freq file
int postingFreq = posting.freq;
if (postingFreq == 1)
// optimize freq=1
freq.WriteVInt(1);
// set low bit of doc num.
else
{
freq.WriteVInt(0); // the document number
freq.WriteVInt(postingFreq); // frequency in doc
}
int lastPosition = 0; // write positions
int[] positions = posting.positions;
for (int j = 0; j < postingFreq; j++)
{
// use delta-encoding
int position = positions[j];
prox.WriteVInt(position - lastPosition);
lastPosition = position;
}
// check to see if we switched to a new field
System.String termField = posting.term.Field();
if (currentField != termField)
{
// changing field - see if there is something to save
currentField = termField;
FieldInfo fi = fieldInfos.FieldInfo(currentField);
if (fi.storeTermVector)
{
if (termVectorWriter == null)
{
termVectorWriter = new TermVectorsWriter(directory, segment, fieldInfos);
termVectorWriter.OpenDocument();
}
termVectorWriter.OpenField(currentField);
}
else if (termVectorWriter != null)
{
termVectorWriter.CloseField();
}
}
if (termVectorWriter != null && termVectorWriter.IsFieldOpen())
{
termVectorWriter.AddTerm(posting.term.Text(), postingFreq, posting.positions, posting.offsets);
}
}
if (termVectorWriter != null)
termVectorWriter.CloseDocument();
}
finally
{
// make an effort to close all streams we can but remember and re-throw
// the first exception encountered in this process
System.IO.IOException keep = null;
if (freq != null)
try
{
freq.Close();
}
catch (System.IO.IOException e)
{
if (keep == null)
keep = e;
}
if (prox != null)
try
{
prox.Close();
}
catch (System.IO.IOException e)
{
if (keep == null)
keep = e;
}
if (tis != null)
try
{
tis.Close();
}
catch (System.IO.IOException e)
{
if (keep == null)
keep = e;
}
if (termVectorWriter != null)
try
{
termVectorWriter.Close();
}
catch (System.IO.IOException e)
{
if (keep == null)
keep = e;
}
if (keep != null)
{
throw new System.IO.IOException(keep.StackTrace);
}
}
}
internal TermInfosWriter(Directory directory, System.String segment, FieldInfos fis, int interval, IState state)
{
Initialize(directory, segment, fis, interval, false, state);
other = new TermInfosWriter(directory, segment, fis, interval, true, state);
other.other = this;
}
public /*internal*/ TermInfosWriter(Directory directory, System.String segment, FieldInfos fis, int interval)
{
Initialize(directory, segment, fis, interval, false);
other = new TermInfosWriter(directory, segment, fis, interval, true);
other.other = this;
}
internal FormatPostingsTermsWriter(SegmentWriteState state, FormatPostingsFieldsWriter parent) : base()
{
this.parent = parent;
termsOut = parent.termsOut;
docsWriter = new FormatPostingsDocsWriter(state, this);
}
/*internal*/
public TermInfosWriter(Directory directory, System.String segment, FieldInfos fis, int interval)
{
Initialize(directory, segment, fis, interval, false);
other = new TermInfosWriter(directory, segment, fis, interval, true);
other.other = this;
}
// TODO: would be nice to factor out morme of this, eg the
// FreqProxFieldMergeState, and code to visit all Fields
// under the same FieldInfo together, up into TermsHash*.
// Other writers would presumably share alot of this...
internal override void flush(IDictionary<object, object> threadsAndFields, DocumentsWriter.FlushState state)
{
// Gather all FieldData's that have postings, across all
// ThreadStates
List<object> allFields = new List<object>();
IEnumerator<KeyValuePair<object, object>> it = threadsAndFields.GetEnumerator();
while (it.MoveNext())
{
KeyValuePair<object, object> entry = (KeyValuePair<object, object>)it.Current;
ICollection<object> fields = (ICollection<object>)entry.Value;
IEnumerator<object> fieldsIt = fields.GetEnumerator();
while (fieldsIt.MoveNext())
{
FreqProxTermsWriterPerField perField = (FreqProxTermsWriterPerField)fieldsIt.Current;
if (perField.termsHashPerField.numPostings > 0)
allFields.Add(perField);
}
}
// Sort by field name
allFields.Sort();
int numAllFields = allFields.Count;
TermInfosWriter termsOut = new TermInfosWriter(state.directory,
state.segmentName,
fieldInfos,
state.docWriter.writer.GetTermIndexInterval());
IndexOutput freqOut = state.directory.CreateOutput(state.SegmentFileName(IndexFileNames.FREQ_EXTENSION));
IndexOutput proxOut;
if (fieldInfos.HasProx())
proxOut = state.directory.CreateOutput(state.SegmentFileName(IndexFileNames.PROX_EXTENSION));
else
proxOut = null;
DefaultSkipListWriter skipListWriter = new DefaultSkipListWriter(termsOut.skipInterval,
termsOut.maxSkipLevels,
state.numDocsInRAM, freqOut, proxOut);
int start = 0;
while (start < numAllFields)
{
FieldInfo fieldInfo = ((FreqProxTermsWriterPerField)allFields[start]).fieldInfo;
string fieldName = fieldInfo.name;
int end = start + 1;
while (end < numAllFields && ((FreqProxTermsWriterPerField)allFields[end]).fieldInfo.name.Equals(fieldName))
end++;
FreqProxTermsWriterPerField[] fields = new FreqProxTermsWriterPerField[end - start];
for (int i = start; i < end; i++)
{
fields[i - start] = (FreqProxTermsWriterPerField)allFields[i];
// Aggregate the storePayload as seen by the same
// field across multiple threads
fieldInfo.storePayloads |= fields[i - start].hasPayloads;
}
// If this field has postings then add them to the
// segment
AppendPostings(state, fields, termsOut, freqOut, proxOut, skipListWriter);
for (int i = 0; i < fields.Length; i++)
{
TermsHashPerField perField = fields[i].termsHashPerField;
int numPostings = perField.numPostings;
perField.reset();
perField.shrinkHash(numPostings);
fields[i].reset();
}
start = end;
}
it = threadsAndFields.GetEnumerator();
while (it.MoveNext())
{
KeyValuePair<object, object> entry = (KeyValuePair<object, object>)it.Current;
FreqProxTermsWriterPerThread perThread = (FreqProxTermsWriterPerThread)entry.Key;
perThread.termsHashPerThread.reset(true);
}
freqOut.Close();
if (proxOut != null)
{
state.flushedFiles[state.SegmentFileName(IndexFileNames.PROX_EXTENSION)] = state.SegmentFileName(IndexFileNames.PROX_EXTENSION);
proxOut.Close();
}
termsOut.Close();
// Record all files we have flushed
state.flushedFiles[state.SegmentFileName(IndexFileNames.FIELD_INFOS_EXTENSION)] = state.SegmentFileName(IndexFileNames.FIELD_INFOS_EXTENSION);
state.flushedFiles[state.SegmentFileName(IndexFileNames.FREQ_EXTENSION)] = state.SegmentFileName(IndexFileNames.FREQ_EXTENSION);
state.flushedFiles[state.SegmentFileName(IndexFileNames.TERMS_EXTENSION)] = state.SegmentFileName(IndexFileNames.TERMS_EXTENSION);
state.flushedFiles[state.SegmentFileName(IndexFileNames.TERMS_INDEX_EXTENSION)] = state.SegmentFileName(IndexFileNames.TERMS_INDEX_EXTENSION);
}
// TODO: would be nice to factor out morme of this, eg the
// FreqProxFieldMergeState, and code to visit all Fields
// under the same FieldInfo together, up into TermsHash*.
// Other writers would presumably share alot of this...
internal override void flush(IDictionary <object, object> threadsAndFields, DocumentsWriter.FlushState state)
{
// Gather all FieldData's that have postings, across all
// ThreadStates
List <object> allFields = new List <object>();
IEnumerator <KeyValuePair <object, object> > it = threadsAndFields.GetEnumerator();
while (it.MoveNext())
{
KeyValuePair <object, object> entry = (KeyValuePair <object, object>)it.Current;
ICollection <object> fields = (ICollection <object>)entry.Value;
IEnumerator <object> fieldsIt = fields.GetEnumerator();
while (fieldsIt.MoveNext())
{
FreqProxTermsWriterPerField perField = (FreqProxTermsWriterPerField)fieldsIt.Current;
if (perField.termsHashPerField.numPostings > 0)
{
allFields.Add(perField);
}
}
}
// Sort by field name
allFields.Sort();
int numAllFields = allFields.Count;
TermInfosWriter termsOut = new TermInfosWriter(state.directory,
state.segmentName,
fieldInfos,
state.docWriter.writer.GetTermIndexInterval());
IndexOutput freqOut = state.directory.CreateOutput(state.SegmentFileName(IndexFileNames.FREQ_EXTENSION));
IndexOutput proxOut;
if (fieldInfos.HasProx())
{
proxOut = state.directory.CreateOutput(state.SegmentFileName(IndexFileNames.PROX_EXTENSION));
}
else
{
proxOut = null;
}
DefaultSkipListWriter skipListWriter = new DefaultSkipListWriter(termsOut.skipInterval,
termsOut.maxSkipLevels,
state.numDocsInRAM, freqOut, proxOut);
int start = 0;
while (start < numAllFields)
{
FieldInfo fieldInfo = ((FreqProxTermsWriterPerField)allFields[start]).fieldInfo;
string fieldName = fieldInfo.name;
int end = start + 1;
while (end < numAllFields && ((FreqProxTermsWriterPerField)allFields[end]).fieldInfo.name.Equals(fieldName))
{
end++;
}
FreqProxTermsWriterPerField[] fields = new FreqProxTermsWriterPerField[end - start];
for (int i = start; i < end; i++)
{
fields[i - start] = (FreqProxTermsWriterPerField)allFields[i];
// Aggregate the storePayload as seen by the same
// field across multiple threads
fieldInfo.storePayloads |= fields[i - start].hasPayloads;
}
// If this field has postings then add them to the
// segment
AppendPostings(state, fields, termsOut, freqOut, proxOut, skipListWriter);
for (int i = 0; i < fields.Length; i++)
{
TermsHashPerField perField = fields[i].termsHashPerField;
int numPostings = perField.numPostings;
perField.reset();
perField.shrinkHash(numPostings);
fields[i].reset();
}
start = end;
}
it = threadsAndFields.GetEnumerator();
while (it.MoveNext())
{
KeyValuePair <object, object> entry = (KeyValuePair <object, object>)it.Current;
FreqProxTermsWriterPerThread perThread = (FreqProxTermsWriterPerThread)entry.Key;
perThread.termsHashPerThread.reset(true);
}
freqOut.Close();
if (proxOut != null)
{
state.flushedFiles[state.SegmentFileName(IndexFileNames.PROX_EXTENSION)] = state.SegmentFileName(IndexFileNames.PROX_EXTENSION);
proxOut.Close();
}
termsOut.Close();
// Record all files we have flushed
state.flushedFiles[state.SegmentFileName(IndexFileNames.FIELD_INFOS_EXTENSION)] = state.SegmentFileName(IndexFileNames.FIELD_INFOS_EXTENSION);
state.flushedFiles[state.SegmentFileName(IndexFileNames.FREQ_EXTENSION)] = state.SegmentFileName(IndexFileNames.FREQ_EXTENSION);
state.flushedFiles[state.SegmentFileName(IndexFileNames.TERMS_EXTENSION)] = state.SegmentFileName(IndexFileNames.TERMS_EXTENSION);
state.flushedFiles[state.SegmentFileName(IndexFileNames.TERMS_INDEX_EXTENSION)] = state.SegmentFileName(IndexFileNames.TERMS_INDEX_EXTENSION);
}
private void MergeTerms()
{
try
{
freqOutput = directory.CreateOutput(segment + ".frq");
proxOutput = directory.CreateOutput(segment + ".prx");
termInfosWriter = new TermInfosWriter(directory, segment, fieldInfos, termIndexInterval);
skipInterval = termInfosWriter.skipInterval;
queue = new SegmentMergeQueue(readers.Count);
MergeTermInfos();
}
finally
{
if (freqOutput != null)
freqOutput.Close();
if (proxOutput != null)
proxOutput.Close();
if (termInfosWriter != null)
termInfosWriter.Close();
if (queue != null)
queue.Close();
}
}
/* Walk through all unique text tokens (Posting
* instances) found in this field and serialize them
* into a single RAM segment. */
void AppendPostings(DocumentsWriter.FlushState flushState,
FreqProxTermsWriterPerField[] fields,
TermInfosWriter termsOut,
IndexOutput freqOut,
IndexOutput proxOut,
DefaultSkipListWriter skipListWriter)
{
int fieldNumber = fields[0].fieldInfo.number;
int numFields = fields.Length;
FreqProxFieldMergeState[] mergeStates = new FreqProxFieldMergeState[numFields];
for (int i = 0; i < numFields; i++)
{
FreqProxFieldMergeState fms = mergeStates[i] = new FreqProxFieldMergeState(fields[i]);
System.Diagnostics.Debug.Assert(fms.field.fieldInfo == fields[0].fieldInfo);
// Should always be true
bool result = fms.nextTerm();
System.Diagnostics.Debug.Assert(result);
}
int skipInterval = termsOut.skipInterval;
bool currentFieldOmitTf = fields[0].fieldInfo.omitTf;
// If current field omits tf then it cannot store
// payloads. We silently drop the payloads in this case:
bool currentFieldStorePayloads = currentFieldOmitTf ? false : fields[0].fieldInfo.storePayloads;
FreqProxFieldMergeState[] termStates = new FreqProxFieldMergeState[numFields];
while (numFields > 0)
{
// Get the next term to merge
termStates[0] = mergeStates[0];
int numToMerge = 1;
for (int i = 1; i < numFields; i++)
{
char[] text = mergeStates[i].text;
int textOffset = mergeStates[i].textOffset;
int cmp = compareText(text, textOffset, termStates[0].text, termStates[0].textOffset);
if (cmp < 0)
{
termStates[0] = mergeStates[i];
numToMerge = 1;
}
else if (cmp == 0)
termStates[numToMerge++] = mergeStates[i];
}
int df = 0;
int lastPayloadLength = -1;
int lastDoc = 0;
char[] text_Renamed = termStates[0].text;
int start = termStates[0].textOffset;
long freqPointer = freqOut.GetFilePointer();
long proxPointer;
if (proxOut != null)
proxPointer = proxOut.GetFilePointer();
else
proxPointer = 0;
skipListWriter.ResetSkip();
// Now termStates has numToMerge FieldMergeStates
// which all share the same term. Now we must
// interleave the docID streams.
while (numToMerge > 0)
{
if ((++df % skipInterval) == 0)
{
skipListWriter.SetSkipData(lastDoc, currentFieldStorePayloads, lastPayloadLength);
skipListWriter.BufferSkip(df);
}
FreqProxFieldMergeState minState = termStates[0];
for (int i = 1; i < numToMerge; i++)
if (termStates[i].docID < minState.docID)
minState = termStates[i];
int doc = minState.docID;
int termDocFreq = minState.termFreq;
System.Diagnostics.Debug.Assert(doc < flushState.numDocsInRAM);
System.Diagnostics.Debug.Assert(doc > lastDoc || df == 1);
ByteSliceReader prox = minState.prox;
// Carefully copy over the prox + payload info,
// changing the format to match Lucene's segment
// format.
if (!currentFieldOmitTf)
{
// omitTf == false so we do write positions & payload
System.Diagnostics.Debug.Assert(proxOut != null);
for (int j = 0; j < termDocFreq; j++)
{
int code = prox.ReadVInt();
if (currentFieldStorePayloads)
{
int payloadLength;
if ((code & 1) != 0)
{
// This position has a payload
payloadLength = prox.ReadVInt();
}
else
payloadLength = 0;
if (payloadLength != lastPayloadLength)
{
proxOut.WriteVInt(code | 1);
proxOut.WriteVInt(payloadLength);
lastPayloadLength = payloadLength;
}
else
proxOut.WriteVInt(code & (~1));
if (payloadLength > 0)
copyBytes(prox, proxOut, payloadLength);
}
else
{
System.Diagnostics.Debug.Assert(0 == (code & 1));
proxOut.WriteVInt(code >> 1);
}
} //End for
int newDocCode = (doc - lastDoc) << 1;
if (1 == termDocFreq)
{
freqOut.WriteVInt(newDocCode | 1);
}
else
{
freqOut.WriteVInt(newDocCode);
freqOut.WriteVInt(termDocFreq);
}
}
else
{
// omitTf==true: we store only the docs, without
// term freq, positions, payloads
freqOut.WriteVInt(doc - lastDoc);
}
lastDoc = doc;
if (!minState.nextDoc())
{
// Remove from termStates
int upto = 0;
for (int i = 0; i < numToMerge; i++)
if (termStates[i] != minState)
termStates[upto++] = termStates[i];
numToMerge--;
System.Diagnostics.Debug.Assert(upto == numToMerge);
// Advance this state to the next term
if (!minState.nextTerm())
{
// OK, no more terms, so remove from mergeStates
// as well
upto = 0;
for (int i = 0; i < numFields; i++)
if (mergeStates[i] != minState)
mergeStates[upto++] = mergeStates[i];
numFields--;
System.Diagnostics.Debug.Assert(upto == numFields);
}
}
}
System.Diagnostics.Debug.Assert(df > 0);
// Done merging this term
long skipPointer = skipListWriter.WriteSkip(freqOut);
// Write term
termInfo.Set(df, freqPointer, proxPointer, (int)(skipPointer - freqPointer));
// TODO: we could do this incrementally
UnicodeUtil.UTF16toUTF8(text_Renamed, start, termsUTF8);
// TODO: we could save O(n) re-scan of the term by
// computing the shared prefix with the last term
// while during the UTF8 encoding
termsOut.Add(fieldNumber,
termsUTF8.result,
termsUTF8.length,
termInfo);
}
}
// FIXME: OG: remove hard-coded file names
public static void Test()
{
System.IO.FileInfo file = new System.IO.FileInfo("words.txt");
System.Console.Out.WriteLine(" reading word file containing " + file.Length + " bytes");
System.DateTime start = System.DateTime.Now;
System.Collections.ArrayList keys = System.Collections.ArrayList.Synchronized(new System.Collections.ArrayList(10));
System.IO.FileStream ws = new System.IO.FileStream(file.FullName, System.IO.FileMode.Open, System.IO.FileAccess.Read);
System.IO.StreamReader wr = new System.IO.StreamReader(new System.IO.StreamReader(ws, System.Text.Encoding.Default).BaseStream, new System.IO.StreamReader(ws, System.Text.Encoding.Default).CurrentEncoding);
for (System.String key = wr.ReadLine(); key != null; key = wr.ReadLine())
{
keys.Add(new Term("word", key));
}
wr.Close();
System.DateTime end = System.DateTime.Now;
System.Console.Out.Write(end.Ticks - start.Ticks);
System.Console.Out.WriteLine(" milliseconds to read " + keys.Count + " words");
start = System.DateTime.Now;
System.Random gen = new System.Random((System.Int32) 1251971);
long fp = (gen.Next() & 0xF) + 1;
long pp = (gen.Next() & 0xF) + 1;
int[] docFreqs = new int[keys.Count];
long[] freqPointers = new long[keys.Count];
long[] proxPointers = new long[keys.Count];
for (int i = 0; i < keys.Count; i++)
{
docFreqs[i] = (gen.Next() & 0xF) + 1;
freqPointers[i] = fp;
proxPointers[i] = pp;
fp += (gen.Next() & 0xF) + 1;
;
pp += (gen.Next() & 0xF) + 1;
;
}
end = System.DateTime.Now;
System.Console.Out.Write(end.Ticks - start.Ticks);
System.Console.Out.WriteLine(" milliseconds to generate values");
start = System.DateTime.Now;
Directory store = FSDirectory.GetDirectory("test.store", true);
FieldInfos fis = new FieldInfos();
TermInfosWriter writer = new TermInfosWriter(store, "words", fis);
fis.Add("word", false);
for (int i = 0; i < keys.Count; i++)
{
writer.Add((Term)keys[i], new TermInfo(docFreqs[i], freqPointers[i], proxPointers[i]));
}
writer.Close();
end = System.DateTime.Now;
System.Console.Out.Write(end.Ticks - start.Ticks);
System.Console.Out.WriteLine(" milliseconds to write table");
System.Console.Out.WriteLine(" table occupies " + store.FileLength("words.tis") + " bytes");
start = System.DateTime.Now;
TermInfosReader reader = new TermInfosReader(store, "words", fis);
end = System.DateTime.Now;
System.Console.Out.Write(end.Ticks - start.Ticks);
System.Console.Out.WriteLine(" milliseconds to open table");
start = System.DateTime.Now;
SegmentTermEnum enumerator = reader.Terms();
for (int i = 0; i < keys.Count; i++)
{
enumerator.Next();
Term key = (Term)keys[i];
if (!key.Equals(enumerator.Term()))
{
throw new System.Exception("wrong term: " + enumerator.Term() + ", expected: " + key + " at " + i);
}
TermInfo ti = enumerator.TermInfo();
if (ti.docFreq != docFreqs[i])
{
throw new System.Exception("wrong value: " + System.Convert.ToString(ti.docFreq, 16) + ", expected: " + System.Convert.ToString(docFreqs[i], 16) + " at " + i);
}
if (ti.freqPointer != freqPointers[i])
{
throw new System.Exception("wrong value: " + System.Convert.ToString(ti.freqPointer, 16) + ", expected: " + System.Convert.ToString(freqPointers[i], 16) + " at " + i);
}
if (ti.proxPointer != proxPointers[i])
{
throw new System.Exception("wrong value: " + System.Convert.ToString(ti.proxPointer, 16) + ", expected: " + System.Convert.ToString(proxPointers[i], 16) + " at " + i);
}
}
end = System.DateTime.Now;
System.Console.Out.Write(end.Ticks - start.Ticks);
System.Console.Out.WriteLine(" milliseconds to iterate over " + keys.Count + " words");
start = System.DateTime.Now;
for (int i = 0; i < keys.Count; i++)
{
Term key = (Term)keys[i];
TermInfo ti = reader.Get(key);
if (ti.docFreq != docFreqs[i])
{
throw new System.Exception("wrong value: " + System.Convert.ToString(ti.docFreq, 16) + ", expected: " + System.Convert.ToString(docFreqs[i], 16) + " at " + i);
}
if (ti.freqPointer != freqPointers[i])
{
throw new System.Exception("wrong value: " + System.Convert.ToString(ti.freqPointer, 16) + ", expected: " + System.Convert.ToString(freqPointers[i], 16) + " at " + i);
}
if (ti.proxPointer != proxPointers[i])
{
throw new System.Exception("wrong value: " + System.Convert.ToString(ti.proxPointer, 16) + ", expected: " + System.Convert.ToString(proxPointers[i], 16) + " at " + i);
}
}
end = System.DateTime.Now;
System.Console.Out.Write((end.Ticks - start.Ticks) / (float)keys.Count);
System.Console.Out.WriteLine(" average milliseconds per lookup");
TermEnum e = reader.Terms(new Term("word", "azz"));
System.Console.Out.WriteLine("Word after azz is " + e.Term().text);
reader.Close();
store.Close();
}
/* Walk through all unique text tokens (Posting
* instances) found in this field and serialize them
* into a single RAM segment. */
void AppendPostings(DocumentsWriter.FlushState flushState,
FreqProxTermsWriterPerField[] fields,
TermInfosWriter termsOut,
IndexOutput freqOut,
IndexOutput proxOut,
DefaultSkipListWriter skipListWriter)
{
int fieldNumber = fields[0].fieldInfo.number;
int numFields = fields.Length;
FreqProxFieldMergeState[] mergeStates = new FreqProxFieldMergeState[numFields];
for (int i = 0; i < numFields; i++)
{
FreqProxFieldMergeState fms = mergeStates[i] = new FreqProxFieldMergeState(fields[i]);
System.Diagnostics.Debug.Assert(fms.field.fieldInfo == fields[0].fieldInfo);
// Should always be true
bool result = fms.nextTerm();
System.Diagnostics.Debug.Assert(result);
}
int skipInterval = termsOut.skipInterval;
bool currentFieldOmitTf = fields[0].fieldInfo.omitTf;
// If current field omits tf then it cannot store
// payloads. We silently drop the payloads in this case:
bool currentFieldStorePayloads = currentFieldOmitTf ? false : fields[0].fieldInfo.storePayloads;
FreqProxFieldMergeState[] termStates = new FreqProxFieldMergeState[numFields];
while (numFields > 0)
{
// Get the next term to merge
termStates[0] = mergeStates[0];
int numToMerge = 1;
for (int i = 1; i < numFields; i++)
{
char[] text = mergeStates[i].text;
int textOffset = mergeStates[i].textOffset;
int cmp = compareText(text, textOffset, termStates[0].text, termStates[0].textOffset);
if (cmp < 0)
{
termStates[0] = mergeStates[i];
numToMerge = 1;
}
else if (cmp == 0)
{
termStates[numToMerge++] = mergeStates[i];
}
}
int df = 0;
int lastPayloadLength = -1;
int lastDoc = 0;
char[] text_Renamed = termStates[0].text;
int start = termStates[0].textOffset;
long freqPointer = freqOut.GetFilePointer();
long proxPointer;
if (proxOut != null)
{
proxPointer = proxOut.GetFilePointer();
}
else
{
proxPointer = 0;
}
skipListWriter.ResetSkip();
// Now termStates has numToMerge FieldMergeStates
// which all share the same term. Now we must
// interleave the docID streams.
while (numToMerge > 0)
{
if ((++df % skipInterval) == 0)
{
skipListWriter.SetSkipData(lastDoc, currentFieldStorePayloads, lastPayloadLength);
skipListWriter.BufferSkip(df);
}
FreqProxFieldMergeState minState = termStates[0];
for (int i = 1; i < numToMerge; i++)
{
if (termStates[i].docID < minState.docID)
{
minState = termStates[i];
}
}
int doc = minState.docID;
int termDocFreq = minState.termFreq;
System.Diagnostics.Debug.Assert(doc < flushState.numDocsInRAM);
System.Diagnostics.Debug.Assert(doc > lastDoc || df == 1);
ByteSliceReader prox = minState.prox;
// Carefully copy over the prox + payload info,
// changing the format to match Lucene's segment
// format.
if (!currentFieldOmitTf)
{
// omitTf == false so we do write positions & payload
System.Diagnostics.Debug.Assert(proxOut != null);
for (int j = 0; j < termDocFreq; j++)
{
int code = prox.ReadVInt();
if (currentFieldStorePayloads)
{
int payloadLength;
if ((code & 1) != 0)
{
// This position has a payload
payloadLength = prox.ReadVInt();
}
else
{
payloadLength = 0;
}
if (payloadLength != lastPayloadLength)
{
proxOut.WriteVInt(code | 1);
proxOut.WriteVInt(payloadLength);
lastPayloadLength = payloadLength;
}
else
{
proxOut.WriteVInt(code & (~1));
}
if (payloadLength > 0)
{
copyBytes(prox, proxOut, payloadLength);
}
}
else
{
System.Diagnostics.Debug.Assert(0 == (code & 1));
proxOut.WriteVInt(code >> 1);
}
} //End for
int newDocCode = (doc - lastDoc) << 1;
if (1 == termDocFreq)
{
freqOut.WriteVInt(newDocCode | 1);
}
else
{
freqOut.WriteVInt(newDocCode);
freqOut.WriteVInt(termDocFreq);
}
}
else
{
// omitTf==true: we store only the docs, without
// term freq, positions, payloads
freqOut.WriteVInt(doc - lastDoc);
}
lastDoc = doc;
if (!minState.nextDoc())
{
// Remove from termStates
int upto = 0;
for (int i = 0; i < numToMerge; i++)
{
if (termStates[i] != minState)
{
termStates[upto++] = termStates[i];
}
}
numToMerge--;
System.Diagnostics.Debug.Assert(upto == numToMerge);
// Advance this state to the next term
if (!minState.nextTerm())
{
// OK, no more terms, so remove from mergeStates
// as well
upto = 0;
for (int i = 0; i < numFields; i++)
{
if (mergeStates[i] != minState)
{
mergeStates[upto++] = mergeStates[i];
}
}
numFields--;
System.Diagnostics.Debug.Assert(upto == numFields);
}
}
}
System.Diagnostics.Debug.Assert(df > 0);
// Done merging this term
long skipPointer = skipListWriter.WriteSkip(freqOut);
// Write term
termInfo.Set(df, freqPointer, proxPointer, (int)(skipPointer - freqPointer));
// TODO: we could do this incrementally
UnicodeUtil.UTF16toUTF8(text_Renamed, start, termsUTF8);
// TODO: we could save O(n) re-scan of the term by
// computing the shared prefix with the last term
// while during the UTF8 encoding
termsOut.Add(fieldNumber,
termsUTF8.result,
termsUTF8.length,
termInfo);
}
}
private void WritePostings(Posting[] postings, System.String segment)
{
IndexOutput freq = null, prox = null;
TermInfosWriter tis = null;
TermVectorsWriter termVectorWriter = null;
try
{
//open files for inverse index storage
freq = directory.CreateOutput(segment + ".frq");
prox = directory.CreateOutput(segment + ".prx");
tis = new TermInfosWriter(directory, segment, fieldInfos, termIndexInterval);
TermInfo ti = new TermInfo();
System.String currentField = null;
for (int i = 0; i < postings.Length; i++)
{
Posting posting = postings[i];
// add an entry to the dictionary with pointers to prox and freq files
ti.Set(1, freq.GetFilePointer(), prox.GetFilePointer(), -1);
tis.Add(posting.term, ti);
// add an entry to the freq file
int postingFreq = posting.freq;
if (postingFreq == 1)
{
// optimize freq=1
freq.WriteVInt(1);
}
// set low bit of doc num.
else
{
freq.WriteVInt(0); // the document number
freq.WriteVInt(postingFreq); // frequency in doc
}
int lastPosition = 0; // write positions
int[] positions = posting.positions;
for (int j = 0; j < postingFreq; j++)
{
// use delta-encoding
int position = positions[j];
prox.WriteVInt(position - lastPosition);
lastPosition = position;
}
// check to see if we switched to a new field
System.String termField = posting.term.Field();
if (currentField != termField)
{
// changing field - see if there is something to save
currentField = termField;
FieldInfo fi = fieldInfos.FieldInfo(currentField);
if (fi.storeTermVector)
{
if (termVectorWriter == null)
{
termVectorWriter = new TermVectorsWriter(directory, segment, fieldInfos);
termVectorWriter.OpenDocument();
}
termVectorWriter.OpenField(currentField);
}
else if (termVectorWriter != null)
{
termVectorWriter.CloseField();
}
}
if (termVectorWriter != null && termVectorWriter.IsFieldOpen())
{
termVectorWriter.AddTerm(posting.term.Text(), postingFreq, posting.positions, posting.offsets);
}
}
if (termVectorWriter != null)
{
termVectorWriter.CloseDocument();
}
}
finally
{
// make an effort to close all streams we can but remember and re-throw
// the first exception encountered in this process
System.IO.IOException keep = null;
if (freq != null)
{
try
{
freq.Close();
}
catch (System.IO.IOException e)
{
if (keep == null)
{
keep = e;
}
}
}
if (prox != null)
{
try
{
prox.Close();
}
catch (System.IO.IOException e)
{
if (keep == null)
{
keep = e;
}
}
}
if (tis != null)
{
try
{
tis.Close();
}
catch (System.IO.IOException e)
{
if (keep == null)
{
keep = e;
}
}
}
if (termVectorWriter != null)
{
try
{
termVectorWriter.Close();
}
catch (System.IO.IOException e)
{
if (keep == null)
{
keep = e;
}
}
}
if (keep != null)
{
throw new System.IO.IOException(keep.StackTrace);
}
}
}
/* Walk through all unique text tokens (Posting
* instances) found in this field and serialize them
* into a single RAM segment. */
internal void AppendPostings(ThreadState.FieldData[] fields, TermInfosWriter termsOut, IndexOutput freqOut, IndexOutput proxOut)
{
int fieldNumber = fields[0].fieldInfo.number;
int numFields = fields.Length;
FieldMergeState[] mergeStates = new FieldMergeState[numFields];
for (int i = 0; i < numFields; i++)
{
FieldMergeState fms = mergeStates[i] = new FieldMergeState();
fms.field = fields[i];
fms.postings = fms.field.SortPostings();
System.Diagnostics.Debug.Assert(fms.field.fieldInfo == fields [0].fieldInfo);
// Should always be true
bool result = fms.NextTerm();
System.Diagnostics.Debug.Assert(result);
}
int skipInterval = termsOut.skipInterval;
currentFieldStorePayloads = fields[0].fieldInfo.storePayloads;
FieldMergeState[] termStates = new FieldMergeState[numFields];
while (numFields > 0)
{
// Get the next term to merge
termStates[0] = mergeStates[0];
int numToMerge = 1;
for (int i = 1; i < numFields; i++)
{
char[] text = mergeStates[i].text;
int textOffset = mergeStates[i].textOffset;
int cmp = CompareText(text, textOffset, termStates[0].text, termStates[0].textOffset);
if (cmp < 0)
{
termStates[0] = mergeStates[i];
numToMerge = 1;
}
else if (cmp == 0)
termStates[numToMerge++] = mergeStates[i];
}
int df = 0;
int lastPayloadLength = - 1;
int lastDoc = 0;
char[] text2 = termStates[0].text;
int start = termStates[0].textOffset;
int pos = start;
while (text2[pos] != 0xffff)
pos++;
long freqPointer = freqOut.GetFilePointer();
long proxPointer = proxOut.GetFilePointer();
skipListWriter.ResetSkip();
// Now termStates has numToMerge FieldMergeStates
// which all share the same term. Now we must
// interleave the docID streams.
while (numToMerge > 0)
{
if ((++df % skipInterval) == 0)
{
skipListWriter.SetSkipData(lastDoc, currentFieldStorePayloads, lastPayloadLength);
skipListWriter.BufferSkip(df);
}
FieldMergeState minState = termStates[0];
for (int i = 1; i < numToMerge; i++)
if (termStates[i].docID < minState.docID)
minState = termStates[i];
int doc = minState.docID;
int termDocFreq = minState.termFreq;
System.Diagnostics.Debug.Assert(doc < numDocsInRAM);
System.Diagnostics.Debug.Assert(doc > lastDoc || df == 1);
int newDocCode = (doc - lastDoc) << 1;
lastDoc = doc;
ByteSliceReader prox = minState.prox;
// Carefully copy over the prox + payload info,
// changing the format to match Lucene's segment
// format.
for (int j = 0; j < termDocFreq; j++)
{
int code = prox.ReadVInt();
if (currentFieldStorePayloads)
{
int payloadLength;
if ((code & 1) != 0)
{
// This position has a payload
payloadLength = prox.ReadVInt();
}
else
payloadLength = 0;
if (payloadLength != lastPayloadLength)
{
proxOut.WriteVInt(code | 1);
proxOut.WriteVInt(payloadLength);
lastPayloadLength = payloadLength;
}
else
proxOut.WriteVInt(code & (~ 1));
if (payloadLength > 0)
CopyBytes(prox, proxOut, payloadLength);
}
else
{
System.Diagnostics.Debug.Assert(0 ==(code & 1));
proxOut.WriteVInt(code >> 1);
}
}
if (1 == termDocFreq)
{
freqOut.WriteVInt(newDocCode | 1);
}
else
{
freqOut.WriteVInt(newDocCode);
freqOut.WriteVInt(termDocFreq);
}
if (!minState.NextDoc())
{
// Remove from termStates
int upto = 0;
for (int i = 0; i < numToMerge; i++)
if (termStates[i] != minState)
termStates[upto++] = termStates[i];
numToMerge--;
System.Diagnostics.Debug.Assert(upto == numToMerge);
// Advance this state to the next term
if (!minState.NextTerm())
{
// OK, no more terms, so remove from mergeStates
// as well
upto = 0;
for (int i = 0; i < numFields; i++)
if (mergeStates[i] != minState)
mergeStates[upto++] = mergeStates[i];
numFields--;
System.Diagnostics.Debug.Assert(upto == numFields);
}
}
}
System.Diagnostics.Debug.Assert(df > 0);
// Done merging this term
long skipPointer = skipListWriter.WriteSkip(freqOut);
// Write term
termInfo.Set(df, freqPointer, proxPointer, (int) (skipPointer - freqPointer));
termsOut.Add(fieldNumber, text2, start, pos - start, termInfo);
}
}
/// <summary>Creates a segment from all Postings in the Postings
/// hashes across all ThreadStates & FieldDatas.
/// </summary>
private System.Collections.IList WriteSegment()
{
System.Diagnostics.Debug.Assert(AllThreadsIdle());
System.Diagnostics.Debug.Assert(nextDocID == numDocsInRAM);
System.String segmentName;
segmentName = segment;
TermInfosWriter termsOut = new TermInfosWriter(directory, segmentName, fieldInfos, writer.GetTermIndexInterval());
IndexOutput freqOut = directory.CreateOutput(segmentName + ".frq");
IndexOutput proxOut = directory.CreateOutput(segmentName + ".prx");
// Gather all FieldData's that have postings, across all
// ThreadStates
System.Collections.ArrayList allFields = new System.Collections.ArrayList();
System.Diagnostics.Debug.Assert(AllThreadsIdle());
for (int i = 0; i < threadStates.Length; i++)
{
ThreadState state = threadStates[i];
state.TrimFields();
int numFields = state.numAllFieldData;
for (int j = 0; j < numFields; j++)
{
ThreadState.FieldData fp = state.allFieldDataArray[j];
if (fp.numPostings > 0)
allFields.Add(fp);
}
}
// Sort by field name
allFields.Sort();
int numAllFields = allFields.Count;
skipListWriter = new DefaultSkipListWriter(termsOut.skipInterval, termsOut.maxSkipLevels, numDocsInRAM, freqOut, proxOut);
int start = 0;
while (start < numAllFields)
{
System.String fieldName = ((ThreadState.FieldData) allFields[start]).fieldInfo.name;
int end = start + 1;
while (end < numAllFields && ((ThreadState.FieldData) allFields[end]).fieldInfo.name.Equals(fieldName))
end++;
ThreadState.FieldData[] fields = new ThreadState.FieldData[end - start];
for (int i = start; i < end; i++)
fields[i - start] = (ThreadState.FieldData) allFields[i];
// If this field has postings then add them to the
// segment
AppendPostings(fields, termsOut, freqOut, proxOut);
for (int i = 0; i < fields.Length; i++)
fields[i].ResetPostingArrays();
start = end;
}
freqOut.Close();
proxOut.Close();
termsOut.Close();
// Record all files we have flushed
System.Collections.IList flushedFiles = new System.Collections.ArrayList();
flushedFiles.Add(SegmentFileName(IndexFileNames.FIELD_INFOS_EXTENSION));
flushedFiles.Add(SegmentFileName(IndexFileNames.FREQ_EXTENSION));
flushedFiles.Add(SegmentFileName(IndexFileNames.PROX_EXTENSION));
flushedFiles.Add(SegmentFileName(IndexFileNames.TERMS_EXTENSION));
flushedFiles.Add(SegmentFileName(IndexFileNames.TERMS_INDEX_EXTENSION));
if (hasNorms)
{
WriteNorms(segmentName, numDocsInRAM);
flushedFiles.Add(SegmentFileName(IndexFileNames.NORMS_EXTENSION));
}
if (infoStream != null)
{
long newSegmentSize = SegmentSize(segmentName);
System.String message = String.Format(nf, " oldRAMSize={0:d} newFlushedSize={1:d} docs/MB={2:f} new/old={3:%}",
new Object[] { numBytesUsed, newSegmentSize, (numDocsInRAM / (newSegmentSize / 1024.0 / 1024.0)), (newSegmentSize / numBytesUsed) });
infoStream.WriteLine(message);
}
ResetPostingsData();
nextDocID = 0;
nextWriteDocID = 0;
numDocsInRAM = 0;
files = null;
// Maybe downsize postingsFreeList array
if (postingsFreeList.Length > 1.5 * postingsFreeCount)
{
int newSize = postingsFreeList.Length;
while (newSize > 1.25 * postingsFreeCount)
{
newSize = (int) (newSize * 0.8);
}
Posting[] newArray = new Posting[newSize];
Array.Copy(postingsFreeList, 0, newArray, 0, postingsFreeCount);
postingsFreeList = newArray;
}
return flushedFiles;
}
private void MergeTerms()
{
try
{
freqOutput = directory.CreateOutput(segment + ".frq");
if (HasProx())
proxOutput = directory.CreateOutput(segment + ".prx");
termInfosWriter = new TermInfosWriter(directory, segment, fieldInfos, termIndexInterval);
skipInterval = termInfosWriter.skipInterval;
maxSkipLevels = termInfosWriter.maxSkipLevels;
skipListWriter = new DefaultSkipListWriter(skipInterval, maxSkipLevels, mergedDocs, freqOutput, proxOutput);
queue = new SegmentMergeQueue(readers.Count);
MergeTermInfos();
}
finally
{
if (freqOutput != null)
freqOutput.Close();
if (proxOutput != null)
proxOutput.Close();
if (termInfosWriter != null)
termInfosWriter.Close();
if (queue != null)
queue.Close();
}
}
// FIXME: OG: remove hard-coded file names
public static void Test()
{
System.IO.FileInfo file = new System.IO.FileInfo("words.txt");
System.Console.Out.WriteLine(" reading word file containing " + file.Length + " bytes");
System.DateTime start = System.DateTime.Now;
System.Collections.ArrayList keys = System.Collections.ArrayList.Synchronized(new System.Collections.ArrayList(10));
System.IO.FileStream ws = new System.IO.FileStream(file.FullName, System.IO.FileMode.Open, System.IO.FileAccess.Read);
System.IO.StreamReader wr = new System.IO.StreamReader(new System.IO.StreamReader(ws, System.Text.Encoding.Default).BaseStream, new System.IO.StreamReader(ws, System.Text.Encoding.Default).CurrentEncoding);
for (System.String key = wr.ReadLine(); key != null; key = wr.ReadLine())
keys.Add(new Term("word", key));
wr.Close();
System.DateTime end = System.DateTime.Now;
System.Console.Out.Write(end.Ticks - start.Ticks);
System.Console.Out.WriteLine(" milliseconds to read " + keys.Count + " words");
start = System.DateTime.Now;
System.Random gen = new System.Random((System.Int32) 1251971);
long fp = (gen.Next() & 0xF) + 1;
long pp = (gen.Next() & 0xF) + 1;
int[] docFreqs = new int[keys.Count];
long[] freqPointers = new long[keys.Count];
long[] proxPointers = new long[keys.Count];
for (int i = 0; i < keys.Count; i++)
{
docFreqs[i] = (gen.Next() & 0xF) + 1;
freqPointers[i] = fp;
proxPointers[i] = pp;
fp += (gen.Next() & 0xF) + 1;
;
pp += (gen.Next() & 0xF) + 1;
;
}
end = System.DateTime.Now;
System.Console.Out.Write(end.Ticks - start.Ticks);
System.Console.Out.WriteLine(" milliseconds to generate values");
start = System.DateTime.Now;
Directory store = FSDirectory.GetDirectory("test.store", true);
FieldInfos fis = new FieldInfos();
TermInfosWriter writer = new TermInfosWriter(store, "words", fis);
fis.Add("word", false);
for (int i = 0; i < keys.Count; i++)
writer.Add((Term) keys[i], new TermInfo(docFreqs[i], freqPointers[i], proxPointers[i]));
writer.Close();
end = System.DateTime.Now;
System.Console.Out.Write(end.Ticks - start.Ticks);
System.Console.Out.WriteLine(" milliseconds to write table");
System.Console.Out.WriteLine(" table occupies " + store.FileLength("words.tis") + " bytes");
start = System.DateTime.Now;
TermInfosReader reader = new TermInfosReader(store, "words", fis);
end = System.DateTime.Now;
System.Console.Out.Write(end.Ticks - start.Ticks);
System.Console.Out.WriteLine(" milliseconds to open table");
start = System.DateTime.Now;
SegmentTermEnum enumerator = reader.Terms();
for (int i = 0; i < keys.Count; i++)
{
enumerator.Next();
Term key = (Term) keys[i];
if (!key.Equals(enumerator.Term()))
{
throw new System.Exception("wrong term: " + enumerator.Term() + ", expected: " + key + " at " + i);
}
TermInfo ti = enumerator.TermInfo();
if (ti.docFreq != docFreqs[i])
throw new System.Exception("wrong value: " + System.Convert.ToString(ti.docFreq, 16) + ", expected: " + System.Convert.ToString(docFreqs[i], 16) + " at " + i);
if (ti.freqPointer != freqPointers[i])
throw new System.Exception("wrong value: " + System.Convert.ToString(ti.freqPointer, 16) + ", expected: " + System.Convert.ToString(freqPointers[i], 16) + " at " + i);
if (ti.proxPointer != proxPointers[i])
throw new System.Exception("wrong value: " + System.Convert.ToString(ti.proxPointer, 16) + ", expected: " + System.Convert.ToString(proxPointers[i], 16) + " at " + i);
}
end = System.DateTime.Now;
System.Console.Out.Write(end.Ticks - start.Ticks);
System.Console.Out.WriteLine(" milliseconds to iterate over " + keys.Count + " words");
start = System.DateTime.Now;
for (int i = 0; i < keys.Count; i++)
{
Term key = (Term) keys[i];
TermInfo ti = reader.Get(key);
if (ti.docFreq != docFreqs[i])
throw new System.Exception("wrong value: " + System.Convert.ToString(ti.docFreq, 16) + ", expected: " + System.Convert.ToString(docFreqs[i], 16) + " at " + i);
if (ti.freqPointer != freqPointers[i])
throw new System.Exception("wrong value: " + System.Convert.ToString(ti.freqPointer, 16) + ", expected: " + System.Convert.ToString(freqPointers[i], 16) + " at " + i);
if (ti.proxPointer != proxPointers[i])
throw new System.Exception("wrong value: " + System.Convert.ToString(ti.proxPointer, 16) + ", expected: " + System.Convert.ToString(proxPointers[i], 16) + " at " + i);
}
end = System.DateTime.Now;
System.Console.Out.Write((end.Ticks - start.Ticks) / (float) keys.Count);
System.Console.Out.WriteLine(" average milliseconds per lookup");
TermEnum e = reader.Terms(new Term("word", "azz"));
System.Console.Out.WriteLine("Word after azz is " + e.Term().text);
reader.Close();
store.Close();
}
