Mikhail Khludnev commented on SOLR-11927:

That json hack is not guaranteed, as far as I know json parsing might strip 
empty arrays giving some settings, ie it's might be broken by next version or 
another parser. Child docs need just made first class citizens in Solr there 
are dozen of open tickets about it already. This one is Won't fix, I believe.   

> Deleting child documents in XML request
> ---------------------------------------
>                 Key: SOLR-11927
>                 URL: https://issues.apache.org/jira/browse/SOLR-11927
>             Project: Solr
>          Issue Type: Bug
>      Security Level: Public(Default Security Level. Issues are Public) 
>    Affects Versions: 7.2
>            Reporter: Chris de Kok
>            Priority: Major
> Currently there are some longstanding issues with deleting updating documents 
> which contain child documents, there is a workaround by always adding the 
> __childDocuments__ property e.g
> curl http://localhost:4003/solr/test/update?commitWithin=3000 -d '
>  [
> {id : 1, _childDocuments_ : []}
> ]'
> Would make sure the child documents are deleted and this works ok.
>  Except in the XML variant.
> <update>
>      <add>
>          <doc>
>              <field name="id">1</field>
>              <doc></doc>
>          </doc>
>      </add>
>  </update>
> It will not work because it will actually try to index it as a child document 
> complaining about missing required fields.

This message was sent by Atlassian JIRA

To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org

Reply via email to