triedate vs date

[Jamie Johnson <je...@gmail.com>]

I was recently walking through schema.xml and noticed triedate vs date
and a note that triedate should be considered instead.  I believe I
understand the basic principle behind triedate but is there an
analysis that exists which shows how much bigger an index would be if
triedate were used vs date?  Also how much faster triedate vs date is?
 Lastly, the default precisionStep is 6, has there been any analysis
about speed/size tradeoffs to show how much faster it is with lower
precisionsSteps (i.e. 2 or 4)?

The only writeups I've seen are the ApacheCon PPT that Uwe Schindler
put together and this
http://www.lucidimagination.com/blog/2009/05/13/exploring-lucene-and-solrs-trierange-capabilities/.
 If there are any other resources available I'd greatly appreciate
being pointed to them.

Re: triedate vs date

[Erick Erickson <er...@gmail.com>]

I'd just try it and see, changing this should be pretty
simple to do a before/after with your data and see
if it's acceptable in your situation...

Best
Erick

On Sun, Feb 5, 2012 at 8:32 AM, Jamie Johnson <je...@gmail.com> wrote:
> I was recently walking through schema.xml and noticed triedate vs date
> and a note that triedate should be considered instead.  I believe I
> understand the basic principle behind triedate but is there an
> analysis that exists which shows how much bigger an index would be if
> triedate were used vs date?  Also how much faster triedate vs date is?
>  Lastly, the default precisionStep is 6, has there been any analysis
> about speed/size tradeoffs to show how much faster it is with lower
> precisionsSteps (i.e. 2 or 4)?
>
> The only writeups I've seen are the ApacheCon PPT that Uwe Schindler
> put together and this
> http://www.lucidimagination.com/blog/2009/05/13/exploring-lucene-and-solrs-trierange-capabilities/.
>  If there are any other resources available I'd greatly appreciate
> being pointed to them.