lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Pete Lewis" <p...@uptima.co.uk>
Subject Re: Wildcard prefix
Date Fri, 27 Jun 2003 17:24:10 GMT
Hi Dave

If you stop searching the index after the first n terms, then on average it
should be much quicker.

Would you be able to come up with a method for wildcard prefix that would
stop after n terms?

Thanks

Pete

----- Original Message -----
From: "Dave Kor" <dave.kor@nexusedge.com>
To: "Lucene Developers List" <lucene-dev@jakarta.apache.org>
Sent: Friday, June 27, 2003 12:29 PM
Subject: RE: Wildcard prefix


> Yes it would be easy to do what you suggested however it still would not
> reduce search times for such queries at all. Afterall the worst case would
> still be that the first n terms are all located at the end of the terms
> list.
>
>
>
> Dave Kor Kian Wei
> Consultant
> Product Engineering
> NexusEdge Technologies Pte. Ltd.
> 6 Aljunied Ave 3, #01-02 (Level 4)
> Singapore 389932
> Tel : (+65)848-2552
> Fax : (+65)747-4536
> Web : www.nexusedge.com
>
> > -----Original Message-----
> > From: Pete Lewis [mailto:pete@uptima.co.uk]
> > Sent: Friday, June 27, 2003 3:56 PM
> > To: Lucene Developers List
> > Subject: Re: Wildcard prefix
> >
> >
> > Hi Dave
> >
> > I was wanting to use this in the terabyte range - so yes SCALEABLE is a
> > must.
> >
> > How easy would it be to pick the first n terms from each index
> > that you were
> > searching, where n would be user configurable?
> >
> > Pete
> >
> > ----- Original Message -----
> > From: "Dave Kor" <dave.kor@nexusedge.com>
> > To: "Lucene Developers List" <lucene-dev@jakarta.apache.org>
> > Sent: Friday, June 27, 2003 8:39 AM
> > Subject: RE: Wildcard prefix
> >
> >
> > >
> > > I wish this can be fixed too, but as the original developer of the
> > wildcard
> > > query and someone who has looked it it from as many angles as
possible,
> > I'ld
> > > have to say that it is quite close to impossible to implement
SCALEABLE
> > > prefix queries with the current lucene engine design. Note that the
> > keyword
> > > here is scaleable. It is very possible and indeed very easy to write a
> > > non-scaleable prefix query (I estimate about 20+ lines of code). The
> > problem
> > > is that this prefix query code's execution speed will slow down
linearly
> > in
> > > proportion to the the number of unique words. In simple english, if
the
> > > search index is small and has less than.. say 50,000 unique terms this
> > > prefix query would run reasonably fast. However when the search index
> > > becomes large and has a few million unique terms, the same prefix
query
> > code
> > > would run at a snail's pace.
> > >
> > > That is why this bug is listed as WONTFIX... no one has come up with a
> > > simple, workable, scaleable solution. sadly I am guessing that
> > other than
> > a
> > > major retrofit to the design of lucene, this issue won't go away.
> > >
> > >
> > >
> > > Dave Kor Kian Wei
> > > Consultant
> > > Product Engineering
> > > NexusEdge Technologies Pte. Ltd.
> > > 6 Aljunied Ave 3, #01-02 (Level 4)
> > > Singapore 389932
> > > Tel : (+65)848-2552
> > > Fax : (+65)747-4536
> > > Web : www.nexusedge.com
> > >
> > > > -----Original Message-----
> > > > From: Pete Lewis [mailto:pete@uptima.co.uk]
> > > > Sent: Friday, June 27, 2003 2:59 PM
> > > > To: Lucene Developers List
> > > > Subject: Wildcard prefix
> > > >
> > > >
> > > > Hi all
> > > >
> > > > The only real functionality that Lucene lacks that is supplied by
> > > > other search engines is the wildcard prefix.  Other features like
> > > > hit mark up can be worked around, but not this one.
> > > >
> > > > I saw that the resolution to the reported problem on this was
> > > > WONTFIX, I was wondering whether this decision could be reviewed.
> > > >
> > > > Pete
> > >
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: lucene-dev-unsubscribe@jakarta.apache.org
> > > For additional commands, e-mail: lucene-dev-help@jakarta.apache.org
> > >
> > >
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: lucene-dev-unsubscribe@jakarta.apache.org
> > For additional commands, e-mail: lucene-dev-help@jakarta.apache.org
> >
> >
> >
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: lucene-dev-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: lucene-dev-help@jakarta.apache.org
>
>


---------------------------------------------------------------------
To unsubscribe, e-mail: lucene-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: lucene-dev-help@jakarta.apache.org


Mime
View raw message