On Sun, Jan 18, 2009 at 11:31 AM, Paul Davis
<[email protected]> wrote:
> The reason that I don't try and index the entire DB is because of the
> arbitrary structure to JSON. Indexing a value that isn't a string
> leads to a whole lot of questions with no good answers. By indexing a
> view I let people pick exactly which parts of their db they want
> indexed.

Woot! This is exactly the kind of FTI that I can see myself using for
my various projects. Thanks Paul!

The main advantage to full-db indexing is that it can be used without
knowing what the schema is. So eg you could provide an FTI web-service
for people's local databases, where they replicate the db to you and
you provide search, without them installing Lucene. But I consider
this an enhancement.

>
> If some can dream up a structured indexing that doesn't suck, I'd be
> glad to take a look. But after all the playing I've done with it I
> just don't really see the structured stuff as being useful.
>
> HTH,
> Paul Davis
>
> On Sun, Jan 18, 2009 at 1:33 PM, Jan Lehnardt <[email protected]> wrote:
>> Yeah, but in case I just want to index the entire DB I don't want to
>> have another index with a copy of all my data.
>>
>> Cheers
>> Jan
>> --
>>
>> On 18 Jan 2009, at 19:29, Justin Cormack wrote:
>>
>>> You can just make a view with everything in if you want to index the
>>> entire
>>> DB.
>>>
>>> Indexing views is much more flexible.
>>>
>>> Justin
>>>
>>> On Sun, Jan 18, 2009 at 1:32 PM, Jan Lehnardt <[email protected]> wrote:
>>>
>>>>
>>>> One thing I noticed (which is different from my original implementation)
>>>> that you do not index databases but only views, is that correct? I'm not
>>>> sure if it matters, but I do see the case for "meh just index the entire
>>>> database". Is this a planned feature?
>>>>
>>>> Cheers
>>>> Jan
>>>> --
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>>
>>>>> Thanks,
>>>>> Paul Davis
>>>>>
>>>>> On Thu, Jan 15, 2009 at 12:50 PM, Paul Davis
>>>>> <[email protected]> wrote:
>>>>>
>>>>>> Darrin,
>>>>>>
>>>>>> action2 went away because _external is now in trunk. My couchdb-lucene
>>>>>> branch is still there and I've put it on the schedule to bring into
>>>>>> conformance with the very slightly altered trunk version (trunk
>>>>>> includes more info in the line API).
>>>>>>
>>>>>> Moving forward you can pull a copy of the couchdb-lucene stuff and
>>>>>> hack on it to get it working or wait for me to get to it this weekend.
>>>>>>
>>>>>> HTH,
>>>>>> Paul Davis
>>>>>>
>>>>>> On Thu, Jan 15, 2009 at 11:09 AM, Darrin Wortlehock
>>>>>> <[email protected]> wrote:
>>>>>>
>>>>>>> Hi,
>>>>>>>
>>>>>>> In late November I was working with davisp's couchdb-lucene, which
>>>>>>> required
>>>>>>> jchris' action2 branch to work.
>>>>>>>
>>>>>>> I am now returning to the development I was doing, building a new
>>>>>>> development environment from scratch and it appears that jchris'
>>>>>>> action2
>>>>>>> branch has now gone away.
>>>>>>>
>>>>>>> I am specifically interested in lucene integration. Can anybody
>>>>>>> provide
>>>>>>> any
>>>>>>> hints on how to move forward?
>>>>>>>
>>>>>>> Regards,
>>>>>>>
>>>>>>> Darrin
>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>
>>
>>
>



-- 
Chris Anderson
http://jchris.mfdz.com

Reply via email to