I had the same problem a while ago; I accidentally forgot to supply the schema
when creating a core and had the digest/string*s* problem. Here's some more
explanation of what I did, I hope I remember correctly:
I am using (just like Markus suggested) the schema.xml from nutch which sets
<fieldType name="string" class="solr.StrField" ...
<field name="digest" type="string" ...
I duplicated the data_driven_schema_configs/conf/ configset and replaced the
managed-schema file with Nutch's schema.xml so I have:
Core is created like so:
solr create -c corename -d my_config
Hope that helps,
> -----Ursprüngliche Nachricht-----
> Von: Markus Jelsma [mailto:markus.jel...@openindex.io]
> Gesendet: Freitag, 14. Oktober 2016 17:05
> An: email@example.com
> Betreff: RE: Nutch 2, Solr 5 - solrdedup causes ClassCastException:
> Yes, current Solr comes with a schemaless mode, which can cause errors in
> some cases. It must be configured to use classic schema mode, then you provide
> it with the schema.xml file you can find in Nutch' conf directory.
> -----Original message-----
> > From:Tom Chiverton <t...@extravision.com>
> > Sent: Friday 14th October 2016 16:58
> > To: firstname.lastname@example.org
> > Subject: Re: Nutch 2, Solr 5 - solrdedup causes ClassCastException:
> > I don't understand what you mean here. I am not a Solr expert, though
> > I've used it a bit in the past, though not with Nutch.
> > Is there a schema I should be feeding it ?
> > Tom
> > On 14/10/16 15:50, Markus Jelsma wrote:
> > > Solr supports schemaless mode, which may be your case. Perhaps it made
> your digest field multi valued. I'd suggest to use Solr's classic schema
> and a fixed schema.