Question

Using Solr 3.5.0 and in my schema.xml I’m using the following to mark the end of sentences and replace the end punctuation with a symbolic token:

<charFilter class="solr.PatternReplaceCharFilterFactory" 
pattern="(?<=[^.!?\\s][^.!?]*(?:[.!?](?![']?\s|$)[^.!?]*)*)[.!?]+(?=\\s|$)"
replacement=" monkeysentence"/>

I’m not sure if that will even work for what I want, but first I need to solve the problem of escaping the ‘<‘ character in the first ‘?<=’ lookbehind.

I get the following error:

org.xml.sax.SAXParseException: The value of attribute "pattern" 
associated with an element type "null" must not contain the '<' character.

I’ve tried using a ‘\’ as in:

 pattern="(?\<=[^.!?\\s][^.!?]*(?:[.!?](?![']?\s|$)[^.!?]*)*)[.!?]+(?=\\s|$)"

But I get the same error.

Answer 1

As this is in an XML file, you will need to use an XML escape to encode <, namely < (you may also need to encode > as >, " as ", and & as &)