/// <summary>Returns a {@link Status} instance detailing
/// the state of the index.
///
/// </summary>
/// <param name="onlySegments">list of specific segment names to check
///
/// <p/>As this method checks every byte in the specified
/// segments, on a large index it can take quite a long
/// time to run.
///
/// <p/><b>WARNING</b>: make sure
/// you only call this when the index is not opened by any
/// writer.
/// </param>
public virtual Status CheckIndex_Renamed_Method(System.Collections.IList onlySegments)
{
System.Globalization.NumberFormatInfo nf = System.Globalization.CultureInfo.CurrentCulture.NumberFormat;
SegmentInfos sis = new SegmentInfos();
Status result = new Status();
result.dir = dir;
try
{
sis.Read(dir);
}
catch (System.Exception t)
{
Msg("ERROR: could not read any segments file in directory");
result.missingSegments = true;
if (infoStream != null)
{
infoStream.WriteLine(t.StackTrace);
}
return(result);
}
int numSegments = sis.Count;
System.String segmentsFileName = sis.GetCurrentSegmentFileName();
IndexInput input = null;
try
{
input = dir.OpenInput(segmentsFileName);
}
catch (System.Exception t)
{
Msg("ERROR: could not open segments file in directory");
if (infoStream != null)
{
infoStream.WriteLine(t.StackTrace);
}
result.cantOpenSegments = true;
return(result);
}
int format = 0;
try
{
format = input.ReadInt();
}
catch (System.Exception t)
{
Msg("ERROR: could not read segment file version in directory");
if (infoStream != null)
{
infoStream.WriteLine(t.StackTrace);
}
result.missingSegmentVersion = true;
return(result);
}
finally
{
if (input != null)
{
input.Close();
}
}
System.String sFormat = "";
bool skip = false;
if (format == SegmentInfos.FORMAT)
{
sFormat = "FORMAT [Lucene Pre-2.1]";
}
if (format == SegmentInfos.FORMAT_LOCKLESS)
{
sFormat = "FORMAT_LOCKLESS [Lucene 2.1]";
}
else if (format == SegmentInfos.FORMAT_SINGLE_NORM_FILE)
{
sFormat = "FORMAT_SINGLE_NORM_FILE [Lucene 2.2]";
}
else if (format == SegmentInfos.FORMAT_SHARED_DOC_STORE)
{
sFormat = "FORMAT_SHARED_DOC_STORE [Lucene 2.3]";
}
else
{
if (format == SegmentInfos.FORMAT_CHECKSUM)
{
sFormat = "FORMAT_CHECKSUM [Lucene 2.4]";
}
else if (format == SegmentInfos.FORMAT_DEL_COUNT)
{
sFormat = "FORMAT_DEL_COUNT [Lucene 2.4]";
}
else if (format == SegmentInfos.FORMAT_HAS_PROX)
{
sFormat = "FORMAT_HAS_PROX [Lucene 2.4]";
}
else if (format == SegmentInfos.FORMAT_USER_DATA)
{
sFormat = "FORMAT_USER_DATA [Lucene 2.9]";
}
else if (format == SegmentInfos.FORMAT_DIAGNOSTICS)
{
sFormat = "FORMAT_DIAGNOSTICS [Lucene 2.9]";
}
else if (format < SegmentInfos.CURRENT_FORMAT)
{
sFormat = "int=" + format + " [newer version of Lucene than this tool]";
skip = true;
}
else
{
sFormat = format + " [Lucene 1.3 or prior]";
}
}
result.segmentsFileName = segmentsFileName;
result.numSegments = numSegments;
result.segmentFormat = sFormat;
result.userData = sis.GetUserData();
System.String userDataString;
if (sis.GetUserData().Count > 0)
{
userDataString = " userData=" + SupportClass.CollectionsHelper.CollectionToString(sis.GetUserData());
}
else
{
userDataString = "";
}
Msg("Segments file=" + segmentsFileName + " numSegments=" + numSegments + " version=" + sFormat + userDataString);
if (onlySegments != null)
{
result.partial = true;
if (infoStream != null)
{
infoStream.Write("\nChecking only these segments:");
}
System.Collections.IEnumerator it = onlySegments.GetEnumerator();
while (it.MoveNext())
{
if (infoStream != null)
{
infoStream.Write(" " + it.Current);
}
}
System.Collections.IEnumerator e = onlySegments.GetEnumerator();
while (e.MoveNext() == true)
{
result.segmentsChecked.Add(e.Current);
}
Msg(":");
}
if (skip)
{
Msg("\nERROR: this index appears to be created by a newer version of Lucene than this tool was compiled on; please re-compile this tool on the matching version of Lucene; exiting");
result.toolOutOfDate = true;
return(result);
}
result.newSegments = (SegmentInfos)sis.Clone();
result.newSegments.Clear();
for (int i = 0; i < numSegments; i++)
{
SegmentInfo info = sis.Info(i);
if (onlySegments != null && !onlySegments.Contains(info.name))
{
continue;
}
Status.SegmentInfoStatus segInfoStat = new Status.SegmentInfoStatus();
result.segmentInfos.Add(segInfoStat);
Msg(" " + (1 + i) + " of " + numSegments + ": name=" + info.name + " docCount=" + info.docCount);
segInfoStat.name = info.name;
segInfoStat.docCount = info.docCount;
int toLoseDocCount = info.docCount;
SegmentReader reader = null;
try
{
Msg(" compound=" + info.GetUseCompoundFile());
segInfoStat.compound = info.GetUseCompoundFile();
Msg(" hasProx=" + info.GetHasProx());
segInfoStat.hasProx = info.GetHasProx();
Msg(" numFiles=" + info.Files().Count);
segInfoStat.numFiles = info.Files().Count;
Msg(System.String.Format(nf, " size (MB)={0:f}", new System.Object[] { (info.SizeInBytes() / (1024.0 * 1024.0)) }));
segInfoStat.sizeMB = info.SizeInBytes() / (1024.0 * 1024.0);
System.Collections.Generic.IDictionary <string, string> diagnostics = info.GetDiagnostics();
segInfoStat.diagnostics = diagnostics;
if (diagnostics.Count > 0)
{
Msg(" diagnostics = " + SupportClass.CollectionsHelper.CollectionToString(diagnostics));
}
int docStoreOffset = info.GetDocStoreOffset();
if (docStoreOffset != -1)
{
Msg(" docStoreOffset=" + docStoreOffset);
segInfoStat.docStoreOffset = docStoreOffset;
Msg(" docStoreSegment=" + info.GetDocStoreSegment());
segInfoStat.docStoreSegment = info.GetDocStoreSegment();
Msg(" docStoreIsCompoundFile=" + info.GetDocStoreIsCompoundFile());
segInfoStat.docStoreCompoundFile = info.GetDocStoreIsCompoundFile();
}
System.String delFileName = info.GetDelFileName();
if (delFileName == null)
{
Msg(" no deletions");
segInfoStat.hasDeletions = false;
}
else
{
Msg(" has deletions [delFileName=" + delFileName + "]");
segInfoStat.hasDeletions = true;
segInfoStat.deletionsFileName = delFileName;
}
if (infoStream != null)
{
infoStream.Write(" test: open reader.........");
}
reader = SegmentReader.Get(info);
segInfoStat.openReaderPassed = true;
int numDocs = reader.NumDocs();
toLoseDocCount = numDocs;
if (reader.HasDeletions())
{
if (reader.deletedDocs.Count() != info.GetDelCount())
{
throw new System.SystemException("delete count mismatch: info=" + info.GetDelCount() + " vs deletedDocs.count()=" + reader.deletedDocs.Count());
}
if (reader.deletedDocs.Count() > reader.MaxDoc())
{
throw new System.SystemException("too many deleted docs: maxDoc()=" + reader.MaxDoc() + " vs deletedDocs.count()=" + reader.deletedDocs.Count());
}
if (info.docCount - numDocs != info.GetDelCount())
{
throw new System.SystemException("delete count mismatch: info=" + info.GetDelCount() + " vs reader=" + (info.docCount - numDocs));
}
segInfoStat.numDeleted = info.docCount - numDocs;
Msg("OK [" + (segInfoStat.numDeleted) + " deleted docs]");
}
else
{
if (info.GetDelCount() != 0)
{
throw new System.SystemException("delete count mismatch: info=" + info.GetDelCount() + " vs reader=" + (info.docCount - numDocs));
}
Msg("OK");
}
if (reader.MaxDoc() != info.docCount)
{
throw new System.SystemException("SegmentReader.maxDoc() " + reader.MaxDoc() + " != SegmentInfos.docCount " + info.docCount);
}
// Test getFieldNames()
if (infoStream != null)
{
infoStream.Write(" test: fields..............");
}
System.Collections.Generic.ICollection <string> fieldNames = reader.GetFieldNames(IndexReader.FieldOption.ALL);
Msg("OK [" + fieldNames.Count + " fields]");
segInfoStat.numFields = fieldNames.Count;
// Test Field Norms
segInfoStat.fieldNormStatus = TestFieldNorms(fieldNames, reader);
// Test the Term Index
segInfoStat.termIndexStatus = TestTermIndex(info, reader);
// Test Stored Fields
segInfoStat.storedFieldStatus = TestStoredFields(info, reader, nf);
// Test Term Vectors
segInfoStat.termVectorStatus = TestTermVectors(info, reader, nf);
// Rethrow the first exception we encountered
// This will cause stats for failed segments to be incremented properly
if (segInfoStat.fieldNormStatus.error != null)
{
throw new System.SystemException("Field Norm test failed");
}
else if (segInfoStat.termIndexStatus.error != null)
{
throw new System.SystemException("Term Index test failed");
}
else if (segInfoStat.storedFieldStatus.error != null)
{
throw new System.SystemException("Stored Field test failed");
}
else if (segInfoStat.termVectorStatus.error != null)
{
throw new System.SystemException("Term Vector test failed");
}
Msg("");
}
catch (System.Exception t)
{
Msg("FAILED");
System.String comment;
comment = "fixIndex() would remove reference to this segment";
Msg(" WARNING: " + comment + "; full exception:");
if (infoStream != null)
{
infoStream.WriteLine(t.StackTrace);
}
Msg("");
result.totLoseDocCount += toLoseDocCount;
result.numBadSegments++;
continue;
}
finally
{
if (reader != null)
{
reader.Close();
}
}
// Keeper
result.newSegments.Add(info.Clone());
}
if (0 == result.numBadSegments)
{
result.clean = true;
Msg("No problems were detected with this index.\n");
}
else
{
Msg("WARNING: " + result.numBadSegments + " broken segments (containing " + result.totLoseDocCount + " documents) detected");
}
return(result);
}
/// <summary>Returns a <see cref="Status" /> instance detailing
/// the state of the index.
///
/// </summary>
/// <param name="onlySegments">list of specific segment names to check
///
/// <p/>As this method checks every byte in the specified
/// segments, on a large index it can take quite a long
/// time to run.
///
/// <p/><b>WARNING</b>: make sure
/// you only call this when the index is not opened by any
/// writer.
/// </param>
public virtual Status CheckIndex_Renamed_Method(List<string> onlySegments)
{
System.Globalization.NumberFormatInfo nf = System.Globalization.CultureInfo.CurrentCulture.NumberFormat;
SegmentInfos sis = new SegmentInfos();
Status result = new Status();
result.dir = dir;
try
{
sis.Read(dir);
}
catch (System.Exception t)
{
Msg("ERROR: could not read any segments file in directory");
result.missingSegments = true;
if (infoStream != null)
infoStream.WriteLine(t.StackTrace);
return result;
}
int numSegments = sis.Count;
var segmentsFileName = sis.GetCurrentSegmentFileName();
IndexInput input = null;
try
{
input = dir.OpenInput(segmentsFileName);
}
catch (System.Exception t)
{
Msg("ERROR: could not open segments file in directory");
if (infoStream != null)
infoStream.WriteLine(t.StackTrace);
result.cantOpenSegments = true;
return result;
}
int format = 0;
try
{
format = input.ReadInt();
}
catch (System.Exception t)
{
Msg("ERROR: could not read segment file version in directory");
if (infoStream != null)
infoStream.WriteLine(t.StackTrace);
result.missingSegmentVersion = true;
return result;
}
finally
{
if (input != null)
input.Close();
}
System.String sFormat = "";
bool skip = false;
if (format == SegmentInfos.FORMAT)
sFormat = "FORMAT [Lucene Pre-2.1]";
if (format == SegmentInfos.FORMAT_LOCKLESS)
sFormat = "FORMAT_LOCKLESS [Lucene 2.1]";
else if (format == SegmentInfos.FORMAT_SINGLE_NORM_FILE)
sFormat = "FORMAT_SINGLE_NORM_FILE [Lucene 2.2]";
else if (format == SegmentInfos.FORMAT_SHARED_DOC_STORE)
sFormat = "FORMAT_SHARED_DOC_STORE [Lucene 2.3]";
else
{
if (format == SegmentInfos.FORMAT_CHECKSUM)
sFormat = "FORMAT_CHECKSUM [Lucene 2.4]";
else if (format == SegmentInfos.FORMAT_DEL_COUNT)
sFormat = "FORMAT_DEL_COUNT [Lucene 2.4]";
else if (format == SegmentInfos.FORMAT_HAS_PROX)
sFormat = "FORMAT_HAS_PROX [Lucene 2.4]";
else if (format == SegmentInfos.FORMAT_USER_DATA)
sFormat = "FORMAT_USER_DATA [Lucene 2.9]";
else if (format == SegmentInfos.FORMAT_DIAGNOSTICS)
sFormat = "FORMAT_DIAGNOSTICS [Lucene 2.9]";
else if (format < SegmentInfos.CURRENT_FORMAT)
{
sFormat = "int=" + format + " [newer version of Lucene than this tool]";
skip = true;
}
else
{
sFormat = format + " [Lucene 1.3 or prior]";
}
}
result.segmentsFileName = segmentsFileName;
result.numSegments = numSegments;
result.segmentFormat = sFormat;
result.userData = sis.UserData;
System.String userDataString;
if (sis.UserData.Count > 0)
{
userDataString = " userData=" + CollectionsHelper.CollectionToString(sis.UserData);
}
else
{
userDataString = "";
}
Msg("Segments file=" + segmentsFileName + " numSegments=" + numSegments + " version=" + sFormat + userDataString);
if (onlySegments != null)
{
result.partial = true;
if (infoStream != null)
infoStream.Write("\nChecking only these segments:");
foreach(string s in onlySegments)
{
if (infoStream != null)
{
infoStream.Write(" " + s);
}
}
result.segmentsChecked.AddRange(onlySegments);
Msg(":");
}
if (skip)
{
Msg("\nERROR: this index appears to be created by a newer version of Lucene than this tool was compiled on; please re-compile this tool on the matching version of Lucene; exiting");
result.toolOutOfDate = true;
return result;
}
result.newSegments = (SegmentInfos) sis.Clone();
result.newSegments.Clear();
for (int i = 0; i < numSegments; i++)
{
SegmentInfo info = sis.Info(i);
if (onlySegments != null && !onlySegments.Contains(info.name))
continue;
var segInfoStat = new Status.SegmentInfoStatus();
result.segmentInfos.Add(segInfoStat);
Msg(" " + (1 + i) + " of " + numSegments + ": name=" + info.name + " docCount=" + info.docCount);
segInfoStat.name = info.name;
segInfoStat.docCount = info.docCount;
int toLoseDocCount = info.docCount;
SegmentReader reader = null;
try
{
Msg(" compound=" + info.GetUseCompoundFile());
segInfoStat.compound = info.GetUseCompoundFile();
Msg(" hasProx=" + info.HasProx);
segInfoStat.hasProx = info.HasProx;
Msg(" numFiles=" + info.Files().Count);
segInfoStat.numFiles = info.Files().Count;
Msg(System.String.Format(nf, " size (MB)={0:f}", new System.Object[] { (info.SizeInBytes() / (1024.0 * 1024.0)) }));
segInfoStat.sizeMB = info.SizeInBytes() / (1024.0 * 1024.0);
IDictionary<string, string> diagnostics = info.Diagnostics;
segInfoStat.diagnostics = diagnostics;
if (diagnostics.Count > 0)
{
Msg(" diagnostics = " + CollectionsHelper.CollectionToString(diagnostics));
}
int docStoreOffset = info.DocStoreOffset;
if (docStoreOffset != - 1)
{
Msg(" docStoreOffset=" + docStoreOffset);
segInfoStat.docStoreOffset = docStoreOffset;
Msg(" docStoreSegment=" + info.DocStoreSegment);
segInfoStat.docStoreSegment = info.DocStoreSegment;
Msg(" docStoreIsCompoundFile=" + info.DocStoreIsCompoundFile);
segInfoStat.docStoreCompoundFile = info.DocStoreIsCompoundFile;
}
System.String delFileName = info.GetDelFileName();
if (delFileName == null)
{
Msg(" no deletions");
segInfoStat.hasDeletions = false;
}
else
{
Msg(" has deletions [delFileName=" + delFileName + "]");
segInfoStat.hasDeletions = true;
segInfoStat.deletionsFileName = delFileName;
}
if (infoStream != null)
infoStream.Write(" test: open reader.........");
reader = SegmentReader.Get(true, info, IndexReader.DEFAULT_TERMS_INDEX_DIVISOR);
segInfoStat.openReaderPassed = true;
int numDocs = reader.NumDocs();
toLoseDocCount = numDocs;
if (reader.HasDeletions)
{
if (reader.deletedDocs.Count() != info.GetDelCount())
{
throw new System.SystemException("delete count mismatch: info=" + info.GetDelCount() + " vs deletedDocs.count()=" + reader.deletedDocs.Count());
}
if (reader.deletedDocs.Count() > reader.MaxDoc)
{
throw new System.SystemException("too many deleted docs: MaxDoc=" + reader.MaxDoc + " vs deletedDocs.count()=" + reader.deletedDocs.Count());
}
if (info.docCount - numDocs != info.GetDelCount())
{
throw new System.SystemException("delete count mismatch: info=" + info.GetDelCount() + " vs reader=" + (info.docCount - numDocs));
}
segInfoStat.numDeleted = info.docCount - numDocs;
Msg("OK [" + (segInfoStat.numDeleted) + " deleted docs]");
}
else
{
if (info.GetDelCount() != 0)
{
throw new System.SystemException("delete count mismatch: info=" + info.GetDelCount() + " vs reader=" + (info.docCount - numDocs));
}
Msg("OK");
}
if (reader.MaxDoc != info.docCount)
throw new System.SystemException("SegmentReader.MaxDoc " + reader.MaxDoc + " != SegmentInfos.docCount " + info.docCount);
// Test getFieldNames()
if (infoStream != null)
{
infoStream.Write(" test: fields..............");
}
ICollection<string> fieldNames = reader.GetFieldNames(IndexReader.FieldOption.ALL);
Msg("OK [" + fieldNames.Count + " fields]");
segInfoStat.numFields = fieldNames.Count;
// Test Field Norms
segInfoStat.fieldNormStatus = TestFieldNorms(fieldNames, reader);
// Test the Term Index
segInfoStat.termIndexStatus = TestTermIndex(info, reader);
// Test Stored Fields
segInfoStat.storedFieldStatus = TestStoredFields(info, reader, nf);
// Test Term Vectors
segInfoStat.termVectorStatus = TestTermVectors(info, reader, nf);
// Rethrow the first exception we encountered
// This will cause stats for failed segments to be incremented properly
if (segInfoStat.fieldNormStatus.error != null)
{
throw new SystemException("Field Norm test failed");
}
else if (segInfoStat.termIndexStatus.error != null)
{
throw new SystemException("Term Index test failed");
}
else if (segInfoStat.storedFieldStatus.error != null)
{
throw new SystemException("Stored Field test failed");
}
else if (segInfoStat.termVectorStatus.error != null)
{
throw new System.SystemException("Term Vector test failed");
}
Msg("");
}
catch (System.Exception t)
{
Msg("FAILED");
const string comment = "fixIndex() would remove reference to this segment";
Msg(" WARNING: " + comment + "; full exception:");
if (infoStream != null)
infoStream.WriteLine(t.StackTrace);
Msg("");
result.totLoseDocCount += toLoseDocCount;
result.numBadSegments++;
continue;
}
finally
{
if (reader != null)
reader.Close();
}
// Keeper
result.newSegments.Add((SegmentInfo)info.Clone());
}
if (0 == result.numBadSegments)
{
result.clean = true;
Msg("No problems were detected with this index.\n");
}
else
Msg("WARNING: " + result.numBadSegments + " broken segments (containing " + result.totLoseDocCount + " documents) detected");
return result;
}
/// <summary>
/// Returns a <seealso cref="Status"/> instance detailing
/// the state of the index.
/// </summary>
/// <param name="onlySegments"> list of specific segment names to check
///
/// <p>As this method checks every byte in the specified
/// segments, on a large index it can take quite a long
/// time to run.
///
/// <p><b>WARNING</b>: make sure
/// you only call this when the index is not opened by any
/// writer. </param>
public virtual Status DoCheckIndex(IList<string> onlySegments)
{
NumberFormatInfo nf = CultureInfo.CurrentCulture.NumberFormat;
SegmentInfos sis = new SegmentInfos();
Status result = new Status();
result.Dir = Dir;
try
{
sis.Read(Dir);
}
catch (Exception t)
{
Msg(infoStream, "ERROR: could not read any segments file in directory");
result.MissingSegments = true;
if (infoStream != null)
{
// LUCENENET NOTE: Some tests rely on the error type being in
// the message. We can't get the error type with StackTrace, we
// need ToString() for that.
infoStream.WriteLine(t.ToString());
//infoStream.WriteLine(t.StackTrace);
}
return result;
}
// find the oldest and newest segment versions
string oldest = Convert.ToString(int.MaxValue), newest = Convert.ToString(int.MinValue);
string oldSegs = null;
bool foundNonNullVersion = false;
IComparer<string> versionComparator = StringHelper.VersionComparator;
foreach (SegmentCommitInfo si in sis.Segments)
{
string version = si.Info.Version;
if (version == null)
{
// pre-3.1 segment
oldSegs = "pre-3.1";
}
else
{
foundNonNullVersion = true;
if (versionComparator.Compare(version, oldest) < 0)
{
oldest = version;
}
if (versionComparator.Compare(version, newest) > 0)
{
newest = version;
}
}
}
int numSegments = sis.Size();
string segmentsFileName = sis.SegmentsFileName;
// note: we only read the format byte (required preamble) here!
IndexInput input = null;
try
{
input = Dir.OpenInput(segmentsFileName, IOContext.READONCE);
}
catch (Exception t)
{
Msg(infoStream, "ERROR: could not open segments file in directory");
if (infoStream != null)
{
// LUCENENET NOTE: Some tests rely on the error type being in
// the message. We can't get the error type with StackTrace, we
// need ToString() for that.
infoStream.WriteLine(t.ToString());
//infoStream.WriteLine(t.StackTrace);
}
result.CantOpenSegments = true;
return result;
}
int format = 0;
try
{
format = input.ReadInt();
}
catch (Exception t)
{
Msg(infoStream, "ERROR: could not read segment file version in directory");
if (infoStream != null)
{
// LUCENENET NOTE: Some tests rely on the error type being in
// the message. We can't get the error type with StackTrace, we
// need ToString() for that.
infoStream.WriteLine(t.ToString());
//infoStream.WriteLine(t.StackTrace);
}
result.MissingSegmentVersion = true;
return result;
}
finally
{
if (input != null)
{
input.Dispose();
}
}
string sFormat = "";
bool skip = false;
result.SegmentsFileName = segmentsFileName;
result.NumSegments = numSegments;
result.UserData = sis.UserData;
string userDataString;
if (sis.UserData.Count > 0)
{
userDataString = " userData=" + sis.UserData;
}
else
{
userDataString = "";
}
string versionString = null;
if (oldSegs != null)
{
if (foundNonNullVersion)
{
versionString = "versions=[" + oldSegs + " .. " + newest + "]";
}
else
{
versionString = "version=" + oldSegs;
}
}
else
{
versionString = oldest.Equals(newest) ? ("version=" + oldest) : ("versions=[" + oldest + " .. " + newest + "]");
}
Msg(infoStream, "Segments file=" + segmentsFileName + " numSegments=" + numSegments + " " + versionString + " format=" + sFormat + userDataString);
if (onlySegments != null)
{
result.Partial = true;
if (infoStream != null)
{
infoStream.Write("\nChecking only these segments:");
foreach (string s in onlySegments)
{
infoStream.Write(" " + s);
}
}
result.SegmentsChecked.AddRange(onlySegments);
Msg(infoStream, ":");
}
if (skip)
{
Msg(infoStream, "\nERROR: this index appears to be created by a newer version of Lucene than this tool was compiled on; please re-compile this tool on the matching version of Lucene; exiting");
result.ToolOutOfDate = true;
return result;
}
result.NewSegments = (SegmentInfos)sis.Clone();
result.NewSegments.Clear();
result.MaxSegmentName = -1;
for (int i = 0; i < numSegments; i++)
{
SegmentCommitInfo info = sis.Info(i);
int segmentName = 0;
try
{
segmentName = int.Parse /*Convert.ToInt32*/(info.Info.Name.Substring(1));
}
catch
{
}
if (segmentName > result.MaxSegmentName)
{
result.MaxSegmentName = segmentName;
}
if (onlySegments != null && !onlySegments.Contains(info.Info.Name))
{
continue;
}
Status.SegmentInfoStatus segInfoStat = new Status.SegmentInfoStatus();
result.SegmentInfos.Add(segInfoStat);
Msg(infoStream, " " + (1 + i) + " of " + numSegments + ": name=" + info.Info.Name + " docCount=" + info.Info.DocCount);
segInfoStat.Name = info.Info.Name;
segInfoStat.DocCount = info.Info.DocCount;
string version = info.Info.Version;
if (info.Info.DocCount <= 0 && version != null && versionComparator.Compare(version, "4.5") >= 0)
{
throw new Exception("illegal number of documents: maxDoc=" + info.Info.DocCount);
}
int toLoseDocCount = info.Info.DocCount;
AtomicReader reader = null;
try
{
Codec codec = info.Info.Codec;
Msg(infoStream, " codec=" + codec);
segInfoStat.Codec = codec;
Msg(infoStream, " compound=" + info.Info.UseCompoundFile);
segInfoStat.Compound = info.Info.UseCompoundFile;
Msg(infoStream, " numFiles=" + info.Files().Count);
segInfoStat.NumFiles = info.Files().Count;
segInfoStat.SizeMB = info.SizeInBytes() / (1024.0 * 1024.0);
if (info.Info.GetAttribute(Lucene3xSegmentInfoFormat.DS_OFFSET_KEY) == null)
{
// don't print size in bytes if its a 3.0 segment with shared docstores
Msg(infoStream, " size (MB)=" + segInfoStat.SizeMB.ToString(nf));
}
IDictionary<string, string> diagnostics = info.Info.Diagnostics;
segInfoStat.Diagnostics = diagnostics;
if (diagnostics.Count > 0)
{
Msg(infoStream, " diagnostics = " + diagnostics);
}
if (!info.HasDeletions())
{
Msg(infoStream, " no deletions");
segInfoStat.HasDeletions = false;
}
else
{
Msg(infoStream, " has deletions [delGen=" + info.DelGen + "]");
segInfoStat.HasDeletions = true;
segInfoStat.DeletionsGen = info.DelGen;
}
if (infoStream != null)
{
infoStream.Write(" test: open reader.........");
}
reader = new SegmentReader(info, DirectoryReader.DEFAULT_TERMS_INDEX_DIVISOR, IOContext.DEFAULT);
Msg(infoStream, "OK");
segInfoStat.OpenReaderPassed = true;
if (infoStream != null)
{
infoStream.Write(" test: check integrity.....");
}
reader.CheckIntegrity();
Msg(infoStream, "OK");
if (infoStream != null)
{
infoStream.Write(" test: check live docs.....");
}
int numDocs = reader.NumDocs;
toLoseDocCount = numDocs;
if (reader.HasDeletions)
{
if (reader.NumDocs != info.Info.DocCount - info.DelCount)
{
throw new Exception("delete count mismatch: info=" + (info.Info.DocCount - info.DelCount) + " vs reader=" + reader.NumDocs);
}
if ((info.Info.DocCount - reader.NumDocs) > reader.MaxDoc)
{
throw new Exception("too many deleted docs: maxDoc()=" + reader.MaxDoc + " vs del count=" + (info.Info.DocCount - reader.NumDocs));
}
if (info.Info.DocCount - numDocs != info.DelCount)
{
throw new Exception("delete count mismatch: info=" + info.DelCount + " vs reader=" + (info.Info.DocCount - numDocs));
}
Bits liveDocs = reader.LiveDocs;
if (liveDocs == null)
{
throw new Exception("segment should have deletions, but liveDocs is null");
}
else
{
int numLive = 0;
for (int j = 0; j < liveDocs.Length(); j++)
{
if (liveDocs.Get(j))
{
numLive++;
}
}
if (numLive != numDocs)
{
throw new Exception("liveDocs count mismatch: info=" + numDocs + ", vs bits=" + numLive);
}
}
segInfoStat.NumDeleted = info.Info.DocCount - numDocs;
Msg(infoStream, "OK [" + (segInfoStat.NumDeleted) + " deleted docs]");
}
else
{
if (info.DelCount != 0)
{
throw new Exception("delete count mismatch: info=" + info.DelCount + " vs reader=" + (info.Info.DocCount - numDocs));
}
Bits liveDocs = reader.LiveDocs;
if (liveDocs != null)
{
// its ok for it to be non-null here, as long as none are set right?
for (int j = 0; j < liveDocs.Length(); j++)
{
if (!liveDocs.Get(j))
{
throw new Exception("liveDocs mismatch: info says no deletions but doc " + j + " is deleted.");
}
}
}
Msg(infoStream, "OK");
}
if (reader.MaxDoc != info.Info.DocCount)
{
throw new Exception("SegmentReader.maxDoc() " + reader.MaxDoc + " != SegmentInfos.docCount " + info.Info.DocCount);
}
// Test getFieldInfos()
if (infoStream != null)
{
infoStream.Write(" test: fields..............");
}
FieldInfos fieldInfos = reader.FieldInfos;
Msg(infoStream, "OK [" + fieldInfos.Size() + " fields]");
segInfoStat.NumFields = fieldInfos.Size();
// Test Field Norms
segInfoStat.FieldNormStatus = TestFieldNorms(reader, infoStream);
// Test the Term Index
segInfoStat.TermIndexStatus = TestPostings(reader, infoStream, Verbose);
// Test Stored Fields
segInfoStat.StoredFieldStatus = TestStoredFields(reader, infoStream);
// Test Term Vectors
segInfoStat.TermVectorStatus = TestTermVectors(reader, infoStream, Verbose, CrossCheckTermVectors_Renamed);
segInfoStat.DocValuesStatus = TestDocValues(reader, infoStream);
// Rethrow the first exception we encountered
// this will cause stats for failed segments to be incremented properly
if (segInfoStat.FieldNormStatus.Error != null)
{
throw new Exception("Field Norm test failed");
}
else if (segInfoStat.TermIndexStatus.Error != null)
{
throw new Exception("Term Index test failed");
}
else if (segInfoStat.StoredFieldStatus.Error != null)
{
throw new Exception("Stored Field test failed");
}
else if (segInfoStat.TermVectorStatus.Error != null)
{
throw new Exception("Term Vector test failed");
}
else if (segInfoStat.DocValuesStatus.Error != null)
{
throw new Exception("DocValues test failed");
}
Msg(infoStream, "");
}
catch (Exception t)
{
Msg(infoStream, "FAILED");
string comment;
comment = "fixIndex() would remove reference to this segment";
Msg(infoStream, " WARNING: " + comment + "; full exception:");
if (infoStream != null)
{
// LUCENENET NOTE: Some tests rely on the error type being in
// the message. We can't get the error type with StackTrace, we
// need ToString() for that.
infoStream.WriteLine(t.ToString());
//infoStream.WriteLine(t.StackTrace);
}
Msg(infoStream, "");
result.TotLoseDocCount += toLoseDocCount;
result.NumBadSegments++;
continue;
}
finally
{
if (reader != null)
{
reader.Dispose();
}
}
// Keeper
result.NewSegments.Add((SegmentCommitInfo)info.Clone());
}
if (0 == result.NumBadSegments)
{
result.Clean = true;
}
else
{
Msg(infoStream, "WARNING: " + result.NumBadSegments + " broken segments (containing " + result.TotLoseDocCount + " documents) detected");
}
if (!(result.ValidCounter = (result.MaxSegmentName < sis.Counter)))
{
result.Clean = false;
result.NewSegments.Counter = result.MaxSegmentName + 1;
Msg(infoStream, "ERROR: Next segment name counter " + sis.Counter + " is not greater than max segment name " + result.MaxSegmentName);
}
if (result.Clean)
{
Msg(infoStream, "No problems were detected with this index.\n");
}
return result;
}
