On Mon, Feb 27, 2012 at 12:36 PM, Steven A Rowe <sar...@syr.edu> wrote:
> Separately, do you know about the "raw" query parser[2]? I'm not sure if
> it would help, but you may be able to use it in alternate solution.
>
And explicitly route to edismax when dismax syntax is detected in the
query? That would make sense, at least from an aesthetic point of view
(aka code readability).
By the way, I'm not sure that edismax interpreting 'wal mart' as 'wal' OR
'mart' is really a bug that should be fixed. It's a counter-intuitive
behavior, for sure, but - per my understanding - edismax is supposed to
treat consecutive words as parts of an OR clause, not as a single phrase.
If what analyzer gets is changed, it would fix some things, but break some
other things.
One small simplification I can think of for your current setup:
> ShingleFilterFactory[1] takes an option called "tokenSeparator" - if you
> set this to the empty string (""), you can eliminate your
> whitespace-stripping filter.
>
Indeed. Thanks for the pointer.
--
Alex Verkhovsky
