manifoldcf-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Karl Wright <daddy...@gmail.com>
Subject Re: Proposal: Not index blank fields in Solr connector
Date Tue, 07 Oct 2014 16:44:23 GMT
Sorry, I should be more clear.  I'm talking about the "Metadata adjuster"
transformation connector, as described here:
http://manifoldcf.apache.org/release/trunk/en_US/end-user-documentation.html#metadataadjuster

The source code is found under connectors/forcedmetadata.

Karl


On Tue, Oct 7, 2014 at 12:38 PM, Karl Wright <daddywri@gmail.com> wrote:

> It just occurred to me that maybe we can put this functionality in the
> Field Mapping transformation connector, controlled by a checkbox.  That
> would make everyone happy I think.
>
> Karl
>
> On Tue, Oct 7, 2014 at 12:35 PM, Alessandro Benedetti <
> benedetti.alex85@gmail.com> wrote:
>
>> I will copy here my consideration :
>>
>> "So you have 2 scenarios for a field : you have content or you don't.
>>
>> Introducing empty field you can have strange behaviours solr sides
>> because from the human point of view there is no difference between an
>> empty or a null field, but for solr yes, and the proper way to model a
>> field that doesn't have content is to not add the field at all to the
>> document.
>>
>> Introducing empty field can lead to inconsistent behaviour mainly sorting
>> side, or when you want to retrieve documents that have value for that field
>> ( in this case indexing empty fields you are introducing false positive for
>> that search)"
>>
>> Curious to see others opinions and discuss !
>>
>>
>> Cheers
>>
>> 2014-10-07 17:28 GMT+01:00 Karl Wright <daddywri@gmail.com>:
>>
>>> Hi folks,
>>>
>>> We're considering, for MCF 2.0, having the Solr connector not index
>>> empty fields -- just skip them entirely.  This is not backwards-compatible
>>> behavior, for a number of reasons:
>>>
>>> (1) If the field whose value is blank is marked as "required" in
>>> solrschema.xml, Solr will reject the document, where before it would have
>>> allowed it to be indexed;
>>> (2) Queries which make a distinction between no field and an empty field
>>> will no longer work the same way.
>>>
>>> Can I request a show of hands (figuratively) of how many people this
>>> change might affect?
>>>
>>> Thanks!
>>> Karl
>>>
>>>
>>>
>>
>>
>> --
>> --------------------------
>>
>> Benedetti Alessandro
>> Visiting card : http://about.me/alessandro_benedetti
>>
>> "Tyger, tyger burning bright
>> In the forests of the night,
>> What immortal hand or eye
>> Could frame thy fearful symmetry?"
>>
>> William Blake - Songs of Experience -1794 England
>>
>
>

Mime
View raw message