18 apr 2007 kl. 18.25 skrev William Mee:
The only way I could get this information *before* adding a document to an index is to create a token stream manually (and then have this happen all over again when the document is indexed). This isn't a satisfying solution.
Why is it not a satisfying solution? Is tokenization really such a burdon to your system? Is your CPU at 100% at all times? Are you really adding that many documents to you index?
One solution would be to add the document in a contrib/memory index, do your secondary analysis, update the document and then add it to your primary index if applicable.
Another solution would be to implement pre-analyzed fields. There is an old outdated patch of mine that does that, <https:// issues.apache.org/jira/browse/LUCENE-580>.
-- karl --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
