Several things:

1> Please don’t use add-unknown…. It’s fine for prototyping, but guesses field 
definitions.

2> the solrocnfig appears to be malformed, I’m surprised it fires up at all. 
This never terminates for instance: <updateHandler 
class="solr.DirectUpdateHandler2”>

3> Have you tried the HtmlStripCharFilterFactory and/or 
HTMLStripFieldUpdateProcessorFactory to remove the markup>? That would be in 
the schema. Another reason I don’t like “field guessing”, it makes treating 
fields specially difficult.

4>  A ScriptUpdateProcessor might be a better way to do some of the 
transformations if regexes don’t prove to be flexible enough.

What I’d do is start over and do exactly one thing at a time. This looks like 
the situation I’ve found myself in far too often, after a while I have 
accumulated enough changes that I’m shooting myself in the foot.

Best,
Erick

> On Feb 21, 2019, at 10:11 AM, Demian Katz <demian.k...@villanova.edu> wrote:
> 
> updateHandler

Reply via email to