Hello Paul,
I came across this yesterday.
http://mail-archives.apache.org/mod_mbox/lucene-java-dev/200504.mbox/[EMAIL
PROTECTED]
My understanding is that by splitting your fields into two indexes and
putting your keyword fields into one and your complicated stuff into the
other then you can update your keyword index in the usual way delete/re-add
without having to update your other index avoiding the re-analyzing.
Regards
Paul I.
Paul Libbrecht <[EMAIL PROTECTED]> wrote on 08/09/2005
10:29:00:
>
> Hi,
>
> some times ago I posted a comment which asking this question (which is
> by no means new) about updating a Lucene document without re-analyzing,
> that is, where we expect the token-streams to be copied into the new
> document and where I intend to change only a few keyword values.
>
> I cannot find this answer anymore, sadly... It was something like "look
> in Nutch" but I didn't find it there, at least.
>
> Could someone repoint me to this bit ? Updating a document this way
> would be a very important optimization for us as we could unload most
> of the content into lucene's document and modify, later, the document
> with results which need to have swallowed the whole repository first.
>
> thank you much.
>
> paul
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
