[SOLR-3185] PatternReplaceCharFilterFactory can't replace with ampersands in index

• Type: Bug
• Status: Closed
• Priority: Minor
• Resolution: Not A Problem
• Affects Version/s: 3.5
• Fix Version/s: 3.5
• Component/s: Schema and Analysis
• Labels: PatternReplaceCharFilter regex
• Assignee: Unassigned
• Reporter: Mike Spencer
• Created: 29/Feb/12 23:09
• Updated: 02/Mar/12 19:36
• Resolved: 02/Mar/12 15:43

Description

Using solr.PatternReplaceCharFilterFactory to replace

A & B

with

A&B

will result in

A&B

being indexed. Query analysis will give the expected result of

A&B

. I examined the index with both standalone Luke and the schema browser field and the index value is incorrect in both tools.

This is the affected charFilter:

<charFilter class="solr.PatternReplaceCharFilterFactory"
                    pattern="(^\w)\s[&amp;]\s(\w)"
                    replacement="$1&amp;$2" />