solr PDF 和自动换行

Question

I am using Solr to index PDF documents. Everything works well, but there is one problem. If a word in a PDF document has been wrapped to another line, then it is indexed as part of the word plus a hyphen. For example, text like this:

We ran to the beach. We heard more guns, then every-

thing was quiet and a flag went up above the trees.

Here the word everything is broken into parts every- and thing . Now if I search for everything , I will not be able to find this document. How to do it right in this case?

Answer 1

With the advice of Abhijit Bashetti and MatsLindh, the problem was solved. In my schema.xml I added the line

<charFilter class="solr.PatternReplaceCharFilterFactory" pattern="\-\n" replacement=""/>

After that, the word wrap does not interfere with the search.