That analysis pipeline does not store empty fields. If the field value is all 
whitespace or empty, the standard tokenizer will not produce any tokens, so the 
rest of the pipeline just doesn’t run. The token stream is empty and nothing is 
stored.

Solr stores terms with positions. If there is no term (empty), then there is 
nothing to have a position. 

It might be possible to do that with a string field, but this is TextField.

wunder
Walter Underwood
wun...@wunderwood.org
http://observer.wunderwood.org/  (my blog)

> On Apr 4, 2024, at 6:38 AM, Carsten Klement <kont...@carsten-klement.de> 
> wrote:
> 
> You're certainly right, but it's currently more important to me that the 
> application runs as before. In version 6.6 the empty fields were saved.
> 
> I currently don't need queries for fields that don't exist.
> 
> <dynamicField name="*_txt"  type="text_general"    indexed="true"  
> stored="true"/>
> 
>  <fieldType name="text_general" class="solr.TextField" 
> positionIncrementGap="100" multiValued="false">
>    <analyzer type="index">
>      <tokenizer name="standard"/>
>      <filter ignoreCase="true" words="stopwords.txt" name="stop"/>
>      <filter name="lowercase"/>
>    </analyzer>
>    <analyzer type="query">
>      <tokenizer name="standard"/>
>      <filter ignoreCase="true" words="stopwords.txt" name="stop"/>
>      <filter ignoreCase="true" synonyms="synonyms.txt" name="synonymGraph" 
> expand="true"/>
>      <filter name="lowercase"/>
>    </analyzer>
>  </fieldType>
> 
> Am 04.04.24, 15:21 schrieb "Gus Heck" <gus.h...@gmail.com 
> <mailto:gus.h...@gmail.com>>:
> 
> 
> Storing a space, whitespace or empty string for a field is generally a bad
> practice. Doing so makes it impossible to query for documents that don't
> contain the field using the normal syntax (i.e. q=*:* -myField:*)
> 
> 
> On Thu, Apr 4, 2024 at 9:09 AM Carsten Klement <kont...@carsten-klement.de 
> <mailto:kont...@carsten-klement.de>>
> wrote:
> 
> 
>> Hi,
>> 
>> we are currently updating to 9.5 (from 6.6) and have the problem that
>> empty fields are not stored in the solr-index.
>> 
>> Is it possible to keep this empty fields?
>> 
>> 
>> 
>> We transmit the documents via http in json structure.
>> 
>> 
>> 
>> Thanks,
>> 
>> Carsten
>> 
>> 
> 
> 
> -- 
> http://www.needhamsoftware.com <http://www.needhamsoftware.com> (work)
> https://a.co/d/b2sZLD9 <https://a.co/d/b2sZLD9> (my fantasy fiction book)
> 
> 
> 

Reply via email to