15 maj 2008 kl. 19.15 skrev Jean-Claude Antonio:
This work perfectly, but for this we need to have a content field as
new Field("content", content, Field.Store.YES, Field.Index.TOKENIZED)
to be able to update the current document which stores the content.
We wish not to store the content as the size of the index gets big.
Is there a way to get the tokenized field value which is not stored?
Or it would be great if we could use a method like
IndexWriter.updateDocuments(Query) but the only one is
IndexWriter.updateDocuments(Term).
Is Lucene your primary data store? That is considered to be a scary
solution according to many people (although I like the idea) and they
would probably recommend you storing a link to the source data rather
than the data it self. An URL, a primary key, or what not.
Do you store the term vectors? It is then a peice of cake. If not you
have to rebuild the document from the inverted index and that is a
very expensive procedure.
LUCENE-1016 can be used to reconstruct token streams from any index.
Luke also does stuff like this. Again, this is expensive.
karl
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
