[ https://issues.apache.org/jira/browse/SOLR-11450?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16203778#comment-16203778 ]
Tim Allison edited comment on SOLR-11450 at 10/13/17 4:19 PM: -------------------------------------------------------------- Ha. Right. Solr does do its own thing. {{FieldTypePluginLoader}} generates a multiterm analyzer in the TextField by subsetting the TokenizerChain's components that are MultitermAware and/or swapping in a KeywordAnalyzer --[here|https://github.com/apache/lucene-solr/blob/branch_6x/solr/core/src/java/org/apache/solr/schema/FieldTypePluginLoader.java#L182] ...just like {{Analyzer.normalize()}} in 7.x :) Then {{SolrQueryParserBase}} has an {{analyzeIfMultiTermText}} [here|https://github.com/apache/lucene-solr/blob/branch_6x/solr/core/src/java/org/apache/solr/parser/SolrQueryParserBase.java#L883], which in turn calls {{TextField}}'s {{analyzeMultiTerm}} with {{TextField}}'s multitermanalyzer that was built back in the {{FieldTypePluginLoader}} above. So, in Solr 6.x, the basic QueryParser relies on the SolrQueryParserBase and all is good. However, the CPQP doesn't extend the SolrQueryParserBase. Two things make this feel like a bug and not a feature in Solr 6.x: 1) multiterm analysis works for the classic query parser in Solr 6.x 2) multiterm analysis works for CPQP for some multiterms (wildcard/reverse wildcard) and range, but not in the other multiterms: prefix, regex and fuzzy. was (Author: talli...@mitre.org): Ha. Right. Solr does do its own thing. {{FieldTypePluginLoader}} generates a multiterm analyzer in the TextField by subsetting the TokenizerChain's components that are MultitermAware and swapping in a KeywordTokenizer --[here|http://example.com] [https://github.com/apache/lucene-solr/blob/branch_6x/solr/core/src/java/org/apache/solr/schema/FieldTypePluginLoader.java#L182] ...just CustomAnalyzer's {{normalize()}} in 7.x :) Then {{SolrQueryParserBase}} has an {{analyzeIfMultiTermText}} [here|https://github.com/apache/lucene-solr/blob/branch_6x/solr/core/src/java/org/apache/solr/parser/SolrQueryParserBase.java#L883], which in turn calls {{TextField}}'s {{analyzeMultiTerm}} with {{TextField}}'s multitermanalyzer that was built back in the {{FieldTypePluginLoader}} above. So, in Solr 6.x, the basic QueryParser relies on the SolrQueryParserBase and all is good. However, the CPQP doesn't extend the SolrQueryParserBase. Two things make this feel like a bug and not a feature in Solr 6.x: 1) multiterm analysis works for the classic query parser in Solr 6.x 2) multiterm analysis works for CPQP for some multiterms (wildcard/reverse wildcard) and range, but not in the other multiterms: prefix, regex and fuzzy. > ComplexPhraseQParserPlugin not running charfilter for some multiterm queries > in 6.x > ------------------------------------------------------------------------------------ > > Key: SOLR-11450 > URL: https://issues.apache.org/jira/browse/SOLR-11450 > Project: Solr > Issue Type: Bug > Security Level: Public(Default Security Level. Issues are Public) > Affects Versions: 6.6.1 > Reporter: Tim Allison > Priority: Minor > Labels: patch-with-test > Attachments: SOLR-11450-unit-test.patch, SOLR-11450.patch > > > On the user list, [~bjarkebm] reported that the charfilter is not being > applied in PrefixQueries in the ComplexPhraseQParserPlugin in 6.x. Bjarke > fixed my proposed unit tests to prove this failure. All appears to work in > 7.x and trunk. If there are plans to release a 6.6.2, let's fold this in. -- This message was sent by Atlassian JIRA (v6.4.14#64029) --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org