Re: arbitrary sorting

[Erik Hatcher]

Thanks for applying that patch, Otis (oh, and making more work for me 
writing about it! :)  I had intended to get to this eventually.

This is indeed a great patch!

	Erik

On Jan 29, 2004, at 6:10 AM, Otis Gospodnetic wrote:

This really was a half-a-line fix to the score method. :)

I'll check in the code after I rename it as Doug had suggested in his
email below.
Tim - thanks for the elegant contribution.  Do you think you'll be able
to add support for sorting on Float and String fields in the near
future?
Otis

--- Doug Cutting <[EMAIL PROTECTED]> wrote:
Tim Jones wrote:
Does anyone have any feedback after looking at the code I submitted
for
sorting results?  Doug and Erik - do you see this as a good way to
go, or do
you have other ideas in mind?  For example, what about creating a
special
kind of "SortingField"?
I like the implementation, requiring an indexed field, but
recommending
against storing or tokenizing.  Most other folks (including myself)
who've done this use a stored field and then iterate over documents
to
fill the cache.  Instead you're able to fill the cache with a
TermDocs,
which is much more efficient.  Good idea!
That said, I've not had a chance to test the code.  But if someone
tests
it, and it works well, I see no reason not to include it.
We might rename it something like IntegerSortedSearcher and also add
classes called FloatSortedSearcher, StringSortedSearcher, etc.  These
could all build on a base class, AbstractSortedSearcher.

Doug


---------------------------------------------------------------------
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]