Hi all,
I'm having an issue with a corrupt document in Lucene. The fields are being built from an XML document and debugging shows them to be correct. But, after calling
indexModifier.addDocument(
doc);
an issue appears in org.apache.lucene.index.FieldsReader#doc(int) whereby the assignment
int fieldNumber = fieldsStream.readVInt();
sets the number of fields to be in the high 4000s (there are actually 18 fields) and the assignment
int fieldNumber = fieldsStream.readVInt();
sets 88 as the field number.
The result of this is a corrupt index that can be searched but when the fields of the hit document are accessed, throws either a null pointer exception or an IOException.
Any help would be massively appreciated!
Thanks,
Steve