In Lucene and Solr, query words that are prefix or wildcard or fuzzy do not go through analyzer processing. This is for well known understood reasons. However, for my data for a field I might want certain processing to occur. Lowercasing has already been identified in
SOLR-219. Another which I address in the attached patch file is the ability to remove characters that meet a supplied regular expression. I've implemented this as part of SolrQueryParser but there should probably be a more thorough plan such as an analyzer chain expressly for the purpose of being applied to prefix, wildcard, and fuzzy queries.