What Erik said ;)

200 fields is not a problem.  Things to watch out for are:
- more index file and thus more open file descriptors if you use non-compound 
Lucene index format and are working with non-optimized indices (on master - 
optimize your index before it gets to slaves)
- slower merging (I think) with more fields (on master, not slave searchers)
- more memory used if lots of fields don't have their norms turned off (i.e. 
are of sub-optimal type)
- more memory used if you sort by lots of fields


Otis
--
Sematext -- http://sematext.com/ -- Lucene - Solr - Nutch


----- Original Message ----
> From: Erik Hatcher <[EMAIL PROTECTED]>
> To: solr-user@lucene.apache.org
> Sent: Friday, April 25, 2008 8:23:27 AM
> Subject: Re: solr performance for documents with hundreds of fields
> 
> That is well within the boundaries of what Solr/Lucene can handle.
> 
> But, of course, it depends on what you're doing with those fields  
> too.  Putting 200 fields into a dismax qf specification, for example,  
> would surely be bad for performance :)   But querying on only a  
> handful of fields or less at a time  - should be no problem.
> 
>     Erik
> 
> 
> 
> On Apr 25, 2008, at 2:24 AM, Umar Shah wrote:
> > I am just wondering,  because having 200 fields seems like too much  
> > (for
> > me),
> > I want to know if people actually have such kind of schemas and how  
> > well
> > they perform.
> >
> >
> >
> >
> > On Thu, Apr 24, 2008 at 5:10 PM, Grant Ingersoll 
> > wrote:
> >
> >> Are you actually seeing performance problems or just wondering if  
> >> there
> >> will be a performance problem?
> >>
> >> -Grant
> >>
> >>
> >> On Apr 24, 2008, at 7:08 AM, Umar Shah wrote:
> >>
> >>  Hi,
> >>>
> >>> I wanted to know what would be the performance of SOLR  for the  
> >>> following
> >>> scenario:
> >>> the documents contain say 200 fields with
> >>> say 100 of the fields (containing numbers)
> >>> and rest containing short strings of 40-50 character length.
> >>> the sparseness of the data can be assumed to be as approximately  
> >>> 50 fields
> >>> missing per document.
> >>>
> >>> any insights?
> >>>
> >>> can a default value of 0 for missing fields change the  
> >>> performance, how?
> >>>
> >>> thanks in anticipation,
> >>> -umar
> >>>
> >>
> >> --------------------------
> >> Grant Ingersoll
> >>
> >> Lucene Helpful Hints:
> >> http://wiki.apache.org/lucene-java/BasicsOfPerformance
> >> http://wiki.apache.org/lucene-java/LuceneFAQ
> >>
> >>
> >>
> >>
> >>
> >>
> >>

Reply via email to