Thank you for the suggestion.

I ended up biting the bullet and re-working my indexing logic. Luckily the
system itself knows what the "current" version of a document is (otherwise
it won't know which one to display to the user) for any given folder.

I was able to get a static method I could call passing in a folder name. The
method returns the file name of the current version for that folder.

Each time I am doing an incremental update if I find that a document from a
folder hasn't changed I make sure it is the current version before moving
on. If it isn't I remove it from the index.

Than when I am creating a new index or adding files to an existing, for each
file, I have to check the file I am adding to ensure it is the current
version for the folder before adding it.

As you can imagine this slows down indexing (creating a new or updating an
existing) but it ensures content from an old version will never show up in a
query.

Luke

----- Original Message ----- 
From: "Yonik Seeley" <[EMAIL PROTECTED]>
To: "Lucene Users List" <[EMAIL PROTECTED]>; "Justin Swanhart"
<[EMAIL PROTECTED]>
Sent: Thursday, November 18, 2004 1:32 PM
Subject: Re: version documents


> This won't fully work.  You still need to delete the
> original out of the lucene index to avoid it showing
> up in searches.
>
> Example:
> myfile v1:  "I want a cat"
> myfile v2:  "I want a dog"
>
> If you change "cat" to "dog" in myfile, and then do a
> search for "cat", you will *only* get v1 and hence the
> sort on version doesn't help.
>
> -Yonik
>
>
> --- Justin Swanhart <[EMAIL PROTECTED]> wrote:
> > Split the filename into "basefilename" and "version"
> > and make each a keyword.
> >
> > Sort your query by version descending, and only use
> > the first
> > "basefile" you encounter.
>
>
> __________________________________________________
> Do You Yahoo!?
> Tired of spam?  Yahoo! Mail has the best spam protection around
> http://mail.yahoo.com
>
> ---------------------------------------------------------------------
> 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]

Reply via email to