IMO, a "reversed word Index" does not work in this case, because he's looking
for a word in the middle (See curi*).
Another idea is to build word chunks and save them in a second index plus docID
of the first index.
e.g. security go to "security ecurity curity ... ity"
This is much faster to
You delete it first using your id:
>writer.deleteDocuments(term);
>
and then re-add it with the same id:
writer.addDocument(d);
>
Please explain:
How looks your document BEFORE you try to delete it? (Which fields has it?)
Greetings Danny