On Apr 12, 2009, at 8:15 AM, Tim Williams wrote:
On Sun, Apr 12, 2009 at 5:56 AM, Erik Hatcher
<e...@ehatchersolutions.com> wrote:
On Apr 11, 2009, at 9:11 PM, Tim Williams wrote:
On Sat, Apr 11, 2009 at 12:25 PM, Erick Erickson
<erickerick...@gmail.com> wrote:
That'll teach me to scan a post. The link I sent you
is still relevant, but wildcards are NOT intended to be used to
concatenate terms. You want a phrase query or a span query
for that. i.e. "A C F"~# where # is the "slop", that is, the number
of other terms allowed to appear between your desired terms.
SpanQueries are constructed programmatically, and PhraseQueries
are produced by the parser.
As I understand it though, there's no way to use the queryparser to
construct an *ordered* phrase query with slop (which is what I think
he's after), right? I gather that'd have to be done manually with a
SpanNearQuery. I'd love to hear that the query parser has syntax
for
this though...
QueryParser does not create any SpanQuery's, but one can subclass
QueryParser and override getFieldQuery() to put in a SpanQuery
instead of a
PhraseQuery.
Thanks Erik, wouldn't this be an all-or-nothing replacement though?
In other words, by creating ordered SpanNearQuery's as the override,
wouldn't he loose the current unordered PhraseQuery+slop
functionality? I haven't seen a way to subclass the QueryParser to
support both (e.g. extend the syntax)?
As always, it depends. If the QueryParser subclass has a switch to
toggle between SpanNearQuery and PhraseQuery it could controlled by
the code which way to go. But yeah, it's not currently possible to
extend the syntax of QueryParser with a subclass. There is a nice new
open issue with a new query parser implementation that is vastly more
flexible - we'll see that come in to Lucene in the near future.
Erik
---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org
For additional commands, e-mail: java-user-h...@lucene.apache.org
