lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ian Vink" <>
Subject Re: Get DocID after Document insert?
Date Thu, 25 Dec 2008 18:14:12 GMT
You just saved me a gazillion hours of work.....

Check out my Lucene enabled
service-to-the-world-hope-we-all-get-along-better site.


On Wed, Dec 24, 2008 at 9:59 AM, Erick Erickson <>wrote:

> You're really, really playing with fire here. Your document IDs may
> change, and in fact *will* change if you delete a document and
> then optimize. Say you index 100 docs, delete number 50 and
> optimize. Documents that originally had IDs 51-100 will now have
> IDs 50-99 and your hierarchy will be messed up.
> And saying "I'll never delete and then optimize" is just asking for
> trouble <G>...
> So just make your own ID that don't change and use *that* to
> maintain your hierarchy. You can get at these pretty efficiently with
> TermDocs/TermEnum and could, say at startup, read your unique
> IDs into a map along with their current Lucene ID if you really care
> that much about speed. This could just be a simple counter and
> you could find the maximum one currently in your index when you
> needed to add more documents to the index using the methods
> I mentioned.
> Best
> Erick
> On Wed, Dec 24, 2008 at 8:35 AM, Ian Vink <> wrote:
> > I am building up an index with documents that are hierarchical in their
> > relationship to each other. After I insert a Document into the index, how
> > do
> > I know its document ID? I need that to pass to the next document as the
> > "ParentID"
> > Ian
> >

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message