Apache
/
lucene
mirror of https://github.com/apache/lucene.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

make CheckIndex a bit more anal about ghost fields 

git-svn-id: https://svn.apache.org/repos/asf/lucene/dev/trunk@1166928 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
Michael McCandless 2011-09-08 22:24:40 +00:00
parent 46985446cf
commit 896afc4c01
1 changed files with 76 additions and 64 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

140
lucene/src/java/org/apache/lucene/index/CheckIndex.java
View File

@ -886,87 +886,99 @@ public class CheckIndex {
} }
final Terms fieldTerms = fields.terms(field); final Terms fieldTerms = fields.terms(field);
if (fieldTerms instanceof BlockTreeTermsReader.FieldReader) { if (fieldTerms == null) {
final BlockTreeTermsReader.Stats stats = ((BlockTreeTermsReader.FieldReader) fieldTerms).computeStats(); // Unusual: the FieldsEnum returned a field but
assert stats != null; // the Terms for that field is null; this should
if (status.blockTreeStats == null) { // only happen if it's a ghost field (field with
status.blockTreeStats = new HashMap<String,BlockTreeTermsReader.Stats>(); // no terms, eg there used to be terms but all
// docs got deleted and then merged away):
// make sure TermsEnum is empty:
if (fieldsEnum.terms().next() != null) {
throw new RuntimeException("Fields.terms(field=" + field + ") returned null yet the field appears to have terms");
}
} else {
if (fieldTerms instanceof BlockTreeTermsReader.FieldReader) {
final BlockTreeTermsReader.Stats stats = ((BlockTreeTermsReader.FieldReader) fieldTerms).computeStats();
assert stats != null;
if (status.blockTreeStats == null) {
status.blockTreeStats = new HashMap<String,BlockTreeTermsReader.Stats>();
}
status.blockTreeStats.put(field, stats);
} }
status.blockTreeStats.put(field, stats);
}
if (sumTotalTermFreq != 0) { if (sumTotalTermFreq != 0) {
final long v = fields.terms(field).getSumTotalTermFreq(); final long v = fields.terms(field).getSumTotalTermFreq();
if (v != -1 && sumTotalTermFreq != v) { if (v != -1 && sumTotalTermFreq != v) {
throw new RuntimeException("sumTotalTermFreq for field " + field + "=" + v + " != recomputed sumTotalTermFreq=" + sumTotalTermFreq); throw new RuntimeException("sumTotalTermFreq for field " + field + "=" + v + " != recomputed sumTotalTermFreq=" + sumTotalTermFreq);
}
} }
}
if (sumDocFreq != 0) { if (sumDocFreq != 0) {
final long v = fields.terms(field).getSumDocFreq(); final long v = fields.terms(field).getSumDocFreq();
if (v != -1 && sumDocFreq != v) { if (v != -1 && sumDocFreq != v) {
throw new RuntimeException("sumDocFreq for field " + field + "=" + v + " != recomputed sumDocFreq=" + sumDocFreq); throw new RuntimeException("sumDocFreq for field " + field + "=" + v + " != recomputed sumDocFreq=" + sumDocFreq);
} }
}
// Test seek to last term:
if (lastTerm != null) {
if (terms.seekCeil(lastTerm) != TermsEnum.SeekStatus.FOUND) {
throw new RuntimeException("seek to last term " + lastTerm + " failed");
} }
is.search(new TermQuery(new Term(field, lastTerm)), 1); // Test seek to last term:
} if (lastTerm != null) {
if (terms.seekCeil(lastTerm) != TermsEnum.SeekStatus.FOUND) {
throw new RuntimeException("seek to last term " + lastTerm + " failed");
}
// Test seeking by ord is.search(new TermQuery(new Term(field, lastTerm)), 1);
if (hasOrd && status.termCount-termCountStart > 0) {
long termCount;
try {
termCount = fields.terms(field).getUniqueTermCount();
} catch (UnsupportedOperationException uoe) {
termCount = -1;
} }
if (termCount != -1 && termCount != status.termCount - termCountStart) { // Test seeking by ord
throw new RuntimeException("termCount mismatch " + termCount + " vs " + (status.termCount - termCountStart)); if (hasOrd && status.termCount-termCountStart > 0) {
} long termCount;
try {
termCount = fields.terms(field).getUniqueTermCount();
} catch (UnsupportedOperationException uoe) {
termCount = -1;
}
int seekCount = (int) Math.min(10000L, termCount); if (termCount != -1 && termCount != status.termCount - termCountStart) {
if (seekCount > 0) { throw new RuntimeException("termCount mismatch " + termCount + " vs " + (status.termCount - termCountStart));
BytesRef[] seekTerms = new BytesRef[seekCount]; }
int seekCount = (int) Math.min(10000L, termCount);
if (seekCount > 0) {
BytesRef[] seekTerms = new BytesRef[seekCount];
// Seek by ord // Seek by ord
for(int i=seekCount-1;i>=0;i--) { for(int i=seekCount-1;i>=0;i--) {
long ord = i*(termCount/seekCount); long ord = i*(termCount/seekCount);
terms.seekExact(ord); terms.seekExact(ord);
seekTerms[i] = new BytesRef(terms.term()); seekTerms[i] = new BytesRef(terms.term());
}
// Seek by term
long totDocCount = 0;
for(int i=seekCount-1;i>=0;i--) {
if (terms.seekCeil(seekTerms[i]) != TermsEnum.SeekStatus.FOUND) {
throw new RuntimeException("seek to existing term " + seekTerms[i] + " failed");
} }
// Seek by term
long totDocCount = 0;
for(int i=seekCount-1;i>=0;i--) {
if (terms.seekCeil(seekTerms[i]) != TermsEnum.SeekStatus.FOUND) {
throw new RuntimeException("seek to existing term " + seekTerms[i] + " failed");
}
docs = terms.docs(liveDocs, docs); docs = terms.docs(liveDocs, docs);
if (docs == null) { if (docs == null) {
throw new RuntimeException("null DocsEnum from to existing term " + seekTerms[i]); throw new RuntimeException("null DocsEnum from to existing term " + seekTerms[i]);
}
while(docs.nextDoc() != DocsEnum.NO_MORE_DOCS) {
totDocCount++;
}
} }
while(docs.nextDoc() != DocsEnum.NO_MORE_DOCS) { // TermQuery
totDocCount++; long totDocCount2 = 0;
for(int i=0;i<seekCount;i++) {
totDocCount2 += is.search(new TermQuery(new Term(field, seekTerms[i])), 1).totalHits;
} }
}
// TermQuery if (totDocCount != totDocCount2) {
long totDocCount2 = 0; throw new RuntimeException("search to seek terms produced wrong number of hits: " + totDocCount + " vs " + totDocCount2);
for(int i=0;i<seekCount;i++) { }
totDocCount2 += is.search(new TermQuery(new Term(field, seekTerms[i])), 1).totalHits;
}
if (totDocCount != totDocCount2) {
throw new RuntimeException("search to seek terms produced wrong number of hits: " + totDocCount + " vs " + totDocCount2);
} }
} }
} }