If 'item' is the unique document level, then this can be done with:
unique id: your own design
searchable text fields:
foo_x:
foo_y:
bar_x:
bar_y:
The query becomes:
foo_x:[100 TO *] AND foo_y:[500 TO *]
Note that to search the other fields with dismax, and foo* with the
standard query parser, you'll need to combine the two with the crazy
multi-parser syntax.
On Fri, Mar 26, 2010 at 10:49 AM, Kumaravel Kandasami
<kumaravel.kandas...@gmail.com> wrote:
> I would represent each "item" element as a document, and each attribute as
> the fields of the document.
>
> if the field names are not known upfront, you could create 'dynamic fields'.
>
>
>
>
> Kumar _/|\_
> www.saisk.com
> ku...@saisk.com
> "making a profound difference with knowledge and creativity..."
>
>
> On Fri, Mar 26, 2010 at 12:37 PM, Phil Messenger <p...@miniweb.tv> wrote:
>
>> Hi,
>>
>> I need to store structured information in an index entry for use when
>> filtering. As XML, this could be expressed as:
>>
>> <item>
>> <some_fields_that_are_searched_using_dismax />
>> <data>
>> <item type="foo" x="100" y="200" />
>> <item type="bar" x="300" y="1000" />
>> </data>
>> </item>
>>
>> I want to be able to *filter* search results according to the data in the
>> "item" tags - eg. show all index entries which match the expression
>> "type=foo && x > 100 & y > 500"
>>
>> Having a multivalued field for type, x and y doesn't seem to work here as
>> I need to maintain the relationship between a type/x/y.
>>
>> I'm not sure how to approach this problem. Is writing a custom field type
>> the
>> preferred approach?
>>
>> thanks,
>>
>> Phil.
>>
>>
>
--
Lance Norskog
goks...@gmail.com
