Re: [Spellcheck] NullPointerException on QueryComponent.mergeIds
James, Sorry for the late response. The shard.qt parameter actually solved my problem ! Thanks Jean-Marc 2013/11/12 Dyer, James james.d...@ingramcontent.com Jean-Marc, This might not solve the particular problem you're having, but to get spellcheck to work properly in a distributed enviornment, be sure to set the shards.qt parameter to the name of your request handler. See http://wiki.apache.org/solr/SpellCheckComponent#Distributed_Search_Support. James Dyer Ingram Content Group (615) 213-4311 -Original Message- From: Jean-Marc Desprez [mailto:jm.desp...@gmail.com] Sent: Tuesday, November 12, 2013 8:57 AM To: solr-user@lucene.apache.org Subject: [Spellcheck] NullPointerException on QueryComponent.mergeIds Hello, I'm following this tutorial : http://wiki.apache.org/solr/SolrCloud with a SolR 4.5.0 I'm at the very first step, only two replica and two shard and I have only *one* document in the index. When I try to get a spellcheck, I have this error : java.lang.NullPointerException at org.apache.solr.handler.component.QueryComponent.mergeIds(QueryComponent.java:843) I do not understand what I'm doing wrong and how I can get an error on mergeIds with only one document in the index (merge this doc with ... ??) Some technical details : URL : http://127.0.0.1:8983/solr/bench/select?shards.qt=ri_spell_fr_FRq=sistemdistrib=true If I set distrib to false, no error. My uniqueKey is indexed and stored : field name=ref type=string indexed=true stored=true multiValued=false / uniqueKeyref/uniqueKey My conf : requestHandler name=ri_spell_fr_FR class=solr.SearchHandler lazy=true lst name=defaults bool name=spellchecktrue/bool str name=spellcheck.onlyMorePopulartrue/str str name=spellcheck.extendedResultstrue/str str name=spellcheck.collateExtendedResultstrue/str str name=spellcheck.maxCollationTries3/str str name=spellcheck.collatetrue/str str name=spellcheck.count5/str str name=spellcheck.dictionaryri_spell_fr_FR/str str name=spellcheck.buildfalse/str /lst arr name=components strspellcheck_fr_FR/str /arr /requestHandler searchComponent name=spellcheck_fr_FR class=solr.SpellCheckComponent str name=queryAnalyzerFieldTypesuggest_fr_FR/str lst name=spellchecker str name=nameri_spell_fr_FR/str str name=fieldspell_fr_FR/str str name=spellcheckIndexDir./spellchecker_fr_FR/str str name=distanceMeasureorg.apache.lucene.search.spell.JaroWinklerDistance/str /lst ... /searchComponent With this URL : http://127.0.0.1:8983/solr/bench/select?qt=ri_spell_fr_FRq=sistem I have no error but the response is empty : responselst name=responseHeaderint name=status0/intint name=QTime1/int/lst/response Thanks Jean-Marc
[Spellcheck] NullPointerException on QueryComponent.mergeIds
Hello, I'm following this tutorial : http://wiki.apache.org/solr/SolrCloud with a SolR 4.5.0 I'm at the very first step, only two replica and two shard and I have only *one* document in the index. When I try to get a spellcheck, I have this error : java.lang.NullPointerException at org.apache.solr.handler.component.QueryComponent.mergeIds(QueryComponent.java:843) I do not understand what I'm doing wrong and how I can get an error on mergeIds with only one document in the index (merge this doc with ... ??) Some technical details : URL : http://127.0.0.1:8983/solr/bench/select?shards.qt=ri_spell_fr_FRq=sistemdistrib=true If I set distrib to false, no error. My uniqueKey is indexed and stored : field name=ref type=string indexed=true stored=true multiValued=false / uniqueKeyref/uniqueKey My conf : requestHandler name=ri_spell_fr_FR class=solr.SearchHandler lazy=true lst name=defaults bool name=spellchecktrue/bool str name=spellcheck.onlyMorePopulartrue/str str name=spellcheck.extendedResultstrue/str str name=spellcheck.collateExtendedResultstrue/str str name=spellcheck.maxCollationTries3/str str name=spellcheck.collatetrue/str str name=spellcheck.count5/str str name=spellcheck.dictionaryri_spell_fr_FR/str str name=spellcheck.buildfalse/str /lst arr name=components strspellcheck_fr_FR/str /arr /requestHandler searchComponent name=spellcheck_fr_FR class=solr.SpellCheckComponent str name=queryAnalyzerFieldTypesuggest_fr_FR/str lst name=spellchecker str name=nameri_spell_fr_FR/str str name=fieldspell_fr_FR/str str name=spellcheckIndexDir./spellchecker_fr_FR/str str name=distanceMeasureorg.apache.lucene.search.spell.JaroWinklerDistance/str /lst ... /searchComponent With this URL : http://127.0.0.1:8983/solr/bench/select?qt=ri_spell_fr_FRq=sistem I have no error but the response is empty : responselst name=responseHeaderint name=status0/intint name=QTime1/int/lst/response Thanks Jean-Marc
SolR document with high number of fields
Hi, I have done some research about SolR document with a very high number of fields. In the mailing list archive there's a thread about this subject which answers my question : http://lucene.472066.n3.nabble.com/Dynamic-fields-performance-question-td476337.html . By the way, this post is a little old and I would like to know if it's still applicable for a recent SolR version (4.5) ? Thanks
Multiple word synonym is not found because of an extra token between words
Hi, Let's say I have this synonyms entry : b c = ok My configuration (index time) : 1. WhitespaceTokenizerFactory 2. WordDelimiterFilterFactory with catenateWords=0 3. SynonymFilterFactory The input : a/b c produce (one line per tokenizer/filter) 0:a/b, 1:c 0:a, 1:b, 2:c 0:a, 1:ok So everything is ok, now if I set catenateWords to 1, the same input produce : 0:a/b, 1:c 0:a, 1:b, 1:ab, 2:c 0:a, 1:b, 1:ab, 2:c The synonym filter doesn't match entry because of the extra token ab between b and c. To my mind the synonym should be triggered when a token b and a token c are separate by one position (which is still the case in the second example). Is there any way to make the second example work ? Jean-Marc