While working with Solr Search encountered one issue.
When a search query is single syllable/word greater than 15 characters we are getting no results.
But if the same search query is shortened to less than 15 characters then we are getting results.
How can we increase the character limit to get search result in both the cases?
Case 1: Greater than 15 Characters
curl -XGET 'http://localhost:8983/solr/techproducts/query?debug=query&q=katanasmartwatch'
Result:
{
"responseHeader": {
"status": 0,
"QTime": 0,
"params": {
"q": "katanasmartwatch",
"debug": "query"
}
},
"response": {
"numFound": 0,
"start": 0,
"docs": []
}
}
Case 2: Less than 15 Characters
curl -XGET 'http://localhost:8983/solr/techproducts/query?debug=query&q=katanasmartwatc'
Result
{
"responseHeader": {
"status": 0,
"QTime": 0,
"params": {
"q": "katanasmartwatc",
"debug": "query"
}
},
"response": {
"numFound": 1,
"start": 0,
"docs": [{...}]
}
}
Related
Any idea for count phrase frequecy , like function termfrequecy with sort asc,desc - and number found for the field ?
Data:
ID / name / surname
/ Test / Test Info
/ Test Info / test
Query:
http://localhostsolr/select?q=name:test info OR surname:test info
I would have expected an answer:
{
"responseHeader": {
"zkConnected": true,
"status": 0,
"QTime": 0,
"params": {
"q": "*.*",
"fl": "id"
}
},
"response": {
"numFound": 2,
"start": 0,
"docs": [
{
"id": "1",
"counterInName": 1,
"counterInSubname": 0,
},
{
"id": "2",
"counterInName": 0,
"counterInSubname": 1,
}
]
}
}
I'm Working with Solr 8.0.0 & i am facing the problem when using the exclude Tag
My solr query looks like below:
http://localhost:8984/solr/HappyDemo202/select?q=*:*&
rows=6&
start=0&
wt=json&
fq={!tag=CATFACET}cat:((desktops))&
fq={!tag=TAGFACET}tag:((cool))&
fq={!tag=Price}Price:[1200 TO 1245]&
json.facet={CatFacet:{type:terms,field:cat,domain:{excludeTags:CATFACET},limit:-1,sort:{count:desc}},TagsFacet:{ type:terms,field:tag,domain:{excludeTags:TAGFACET},limit:-1,sort:{count:desc}}}
Output Of Query looks like below:
{ "responseHeader": {
"status": 0,
"QTime": 0,
"params": {
"q": "*:*",
"json.facet": "{CatFacet:{type:terms,field:cat,domain:{excludeTags:CATFACET},limit:-1,sort:{count:desc}},TagsFacet:{ type:terms,field:tag,domain:{excludeTags:TAGFACET},limit:-1,sort:{count:desc}}}",
"start": "0",
"fq": [
"{!tag=CATFACET}cat:((desktops))",
"{!tag=TAGFACET}tag:((cool))",
"{!tag=Price}Price:[1200 TO 1245]"
],
"rows": "6",
"wt": "json"
}
}, "response": {
"numFound": 0,
"start": 0,
"docs": [] },
"facets": {
"count": 0,
"CatFacet": {
"buckets": []
},
"TagsFacet": {
"buckets": [
{
"val": "new",
"count": 1
},
{
"val": "new1",
"count": 1
}
]
} } }
When you check the Output of Query,CatFacet is not showing any facet result because numFound is 0 but TagsFacet is showing the two facet result like new & new1. I don't know what going wrong , tagFacet must not show the two facet result if numFound is 0.
Can you please suggest,what's going wrong ? Any help will be appreciable.
You're explicitly asking for the fq for tag to be excluded for the tag facet ({excludeTags:TAGFACET}) - meaning that if you didn't have that fq there, there would be results - and you're asking for a count without that field.
If you want the facet to only count those documents being returned, drop the excludeTags value for all those facets that should only be returned for documents that are included in the result set.
Solr is removing the letter 'e' from search queries...
I'm pretty new when it comes to Solr so I don't really know where to start looking to figure this out but whenever I send a search query Solr is stripping out the 'e' character...
As you can see here when I try and search the term ace001
{
"responseHeader": {
"status": 0,
"QTime": 1,
"params": {
"q": "_text:ace001",
"indent": "true",
"wt": "json",
"debugQuery": "true",
"_": "1478467316690"
}
},
"response": {
"numFound": 0,
"start": 0,
"docs": []
},
"debug": {
"rawquerystring": "_text:ace001",
"querystring": "_text:ace001",
"parsedquery": "PhraseQuery(_text:\"ac 001 ac 001\")",
"parsedquery_toString": "_text:\"ac 001 ac 001\"",
"explain": {},
"QParser": "LuceneQParser",
"timing": {
"time": 1,
"prepare": {
"time": 1,
"query": {
"time": 1
},
"facet": {
"time": 0
},
"mlt": {
"time": 0
},
"highlight": {
"time": 0
},
"stats": {
"time": 0
},
"spellcheck": {
"time": 0
},
"debug": {
"time": 0
}
},
"process": {
"time": 0,
"query": {
"time": 0
},
"facet": {
"time": 0
},
"mlt": {
"time": 0
},
"highlight": {
"time": 0
},
"stats": {
"time": 0
},
"spellcheck": {
"time": 0
},
"debug": {
"time": 0
}
}
}
}
}
Searching a different term such as 'acb001' doesn't strip the 'b' but I noticed it does separate the numbers from the letters. I'd want Solr to match the term 'acb001' in the text field...
extract:
"rawquerystring": "_text:acb001",
"querystring": "_text:acb001",
"parsedquery": "PhraseQuery(_text:\"acb 001 acb 001\")",
"parsedquery_toString": "_text:\"acb 001 acb 001\"",
"explain": {},
"QParser": "LuceneQParser",
Would really appreciate some direction here as to how I can either further debug or ideally fix this so ace001 returns all the occurrences of just that.
Edit:
Schema is standard/default http://pastebin.com/59LbmJUp
this is happening because of solr.PorterStemFilterFactory. your default search field id is htmltext which has
<filter class="solr.PorterStemFilterFactory"/>
in the query analysis.
the PorterStemmer stems the word "ace" to "ac".
you can check it here https://tartarus.org/martin/PorterStemmer/voc.txt
search for the word "ace".
now look here which has corresponding output after stemming https://tartarus.org/martin/PorterStemmer/output.txt the corresponding word after stemming which will be "ac"
to solve this revmoe the filter during query as well as index in solrconfig.xml
Also you are using WordDelimiterFilterFactory, which will split words on alphanumeric bounderies. that is why you see "ac" and "001", if you do not want that then remove that filter too in schema.xml
you are using default schema.xml which has a lot of these unnecessary filters which you might not even need. I would suggest to strip it down to a few filters. and then add filters as you need instead of the other way.
I have solr 5.2.1 and such definition of field which is used for sorting:
<fieldType name="polishSortVarchar" class="solr.ICUCollationField" locale="pl_PL" strength="secondary" />
After reindex sorting almost work as I want:
{
"responseHeader": {
"status": 0,
"QTime": 2,
"params": {
"fl": "name_varchar",
"sort": "sort_name_varchar asc",
"indent": "true",
"q": "*:*",
"_": "1454575147254",
"wt": "json",
"rows": "10"
}
},
"response": {
"numFound": 5250,
"start": 0,
"docs": [
{
"name_varchar": "\"Europą\" na Antarktydę"
},
{
"name_varchar": "1:0 dla Korniszonka"
},
{
"name_varchar": "1001 faktów o roślinach"
}
]
}
}
As You see on first position is phrase with " on 1st char, I want filter special chars and sort only by letters (so this phrase will be sorted by 'E' on first position).
Anybody?
I can't find solution directly in SOLR, so I clean unnecessary chars during indexation.
$sortValue = preg_replace('/[^A-Za-z0-9- zżźćńółęąśŻŹĆĄŚĘŁÓŃ]/u', '', $sortValue);
When querying solr with a group-by field, I a response with "num_found" greater than 1, yet the "docs" attribute only shows 1 record.
The query is something like:
http://.../solr/.../select?q=*%3A*&fq=...&wt=json&indent=true&group=true&group.field=GroupingField_s&group.ngroups=true
The results are something like:
"grouped": {
"GroupingField_s": {
"matches": 3130,
"ngroups": 283,
"groups": [
{
"groupValue": "1111",
"doclist": {
"numFound": 7,
"start": 0,
"docs": [ {/*only 1 record shown here*/} ]
},
{
"groupValue": "222",
"doclist": {
"numFound": 5,
"start": 0,
"docs": [ {/*only 1 record shown here*/} ]
}, ....
]
}
You'll have to set the group.limit parameter. This defaults to 1.
group.limit integer Specifies the number of results to return for each group. The default value is 1.
See Result Grouping.