bq: metatag.description field has highest boost and documents with a
match in this field should come first
This is something of a misunderstanding. The boost you've added to the
metatag.description field is a _factor_ in the score calculations, not
an ordering. In your case you're using edismax apparently (inferred by
the "max of", you didn't say) which scores based on the _maximum_ of
all the clauses and it happens that by the time you score the "title"
field in the first doc it has a higher score than the
metatag.description field in the second.... My guess is that the title
field in the first doc is much shorter than the title field in the
second...
And as an aside, periods aren't recommended in field names. They work
as far as we know, but going forward aren't officially supported.
Best,
Erick
On Mon, Jun 20, 2016 at 7:07 AM, Megha Bhandari <mbhanda...@sapient.com> wrote:
> Hi
>
> Problem statement : metatag.description field has highest boost and documents
> with a match in this field should come first. However
> Silver-Choice-5000-E.pdf comes before /contact-us/florida even though the
> search term matches more fields in the /contact-us/florida page. In
> Silver-Choice-5000-E.pdf matches are found in title and _text fields. In
> /contact-us/florida matches are found in metatag.description,title,h1,_text
> fields.
>
> We have the following in solrconfig.xml
>
> <requestHandler name="/elevate" class="solr.SearchHandler" startup="lazy">
> <lst name="defaults">
> <str name="defType">edismax</str>
> <str name="qf">
> metatag.keywords^900000.1 metatag.description^500000.1 title^10000.1
> h1^1000.7 h2^700.6 h3^10.1 h4^5.4 h5^1.3 h6^1.2 _text_^1.0
> </str>
>
> When searching for Florida we get the following results.
> ------------------------------------------
>
> {
>
>
> 'id'=>'/content/dam/uhcdotcom/en/qa_workarea/Silver-Choice-5000-E.pdf',
>
> 'title'=>'Florida',
>
> 'metatag.description'=>'mental health',
>
> 'itemtype'=>'pdf',
>
> 'playerid'=>'',
>
> 'playerkey'=>'',
>
> 'metatag.topresultthumbnailalt'=>'Florida',
>
> 'lang'=>'en',
>
> 'metatag.hideininternalsearch'=>'false'},
>
> {
>
> 'lang'=>'en',
>
> 'metatag.topresultthumbnailurl'=>'',
>
> 'id'=>'https://10.209.5.171/contact-us/florida',
>
> 'title'=>'Florida',
>
> 'metatag.topresultthumbnailalt'=>'',
>
> 'metatag.hideininternalsearch'=>'false',
>
> 'metatag.description'=>'Contact UnitedHealthcare in Florida.'}
>
> -----------------------------------------------------------
> With following debug information
>
> '/content/dam/uhcdotcom/en/qa_workarea/Silver-Choice-5000-E.pdf'=>'
> 0.050655644 = max of:
> 0.050655644 = weight(title:florida in 0) [ClassicSimilarity], result of:
> 0.050655644 = score(doc=0,freq=1.0), product of:
> 0.0075930804 = queryWeight, product of:
> 10000.1 = boost
> 6.6712904 = idf(docFreq=21, maxDocs=6389)
> 1.1381613E-7 = queryNorm
> 6.6712904 = fieldWeight in 0, product of:
> 1.0 = tf(freq=1.0), with freq of:
> 1.0 = termFreq=1.0
> 6.6712904 = idf(docFreq=21, maxDocs=6389)
> 1.0 = fieldNorm(doc=0)
> 3.462133E-7 = weight(_text_:florida in 0) [ClassicSimilarity], result of:
> 3.462133E-7 = score(doc=0,freq=2.0), product of:
> 4.222856E-7 = queryWeight, product of:
> 3.710244 = idf(docFreq=424, maxDocs=6389)
> 1.1381613E-7 = queryNorm
> 0.8198558 = fieldWeight in 0, product of:
> 1.4142135 = tf(freq=2.0), with freq of:
> 2.0 = termFreq=2.0
> 3.710244 = idf(docFreq=424, maxDocs=6389)
> 0.15625 = fieldNorm(doc=0)
> ',
> 'https://10.209.5.171/contact-us/florida'=>'
> 0.02968075 = max of:
> 0.0011872416 = weight(title:florida in 380) [ClassicSimilarity], result of:
> 0.0011872416 = score(doc=380,freq=1.0), product of:
> 0.0075930804 = queryWeight, product of:
> 10000.1 = boost
> 6.6712904 = idf(docFreq=21, maxDocs=6389)
> 1.1381613E-7 = queryNorm
> 0.15635836 = fieldWeight in 380, product of:
> 1.0 = tf(freq=1.0), with freq of:
> 1.0 = termFreq=1.0
> 6.6712904 = idf(docFreq=21, maxDocs=6389)
> 0.0234375 = fieldNorm(doc=380)
> 5.724965E-9 = weight(_text_:florida in 380) [ClassicSimilarity], result of:
> 5.724965E-9 = score(doc=380,freq=14.0), product of:
> 4.222856E-7 = queryWeight, product of:
> 3.710244 = idf(docFreq=424, maxDocs=6389)
> 1.1381613E-7 = queryNorm
> 0.013557092 = fieldWeight in 380, product of:
> 3.7416575 = tf(freq=14.0), with freq of:
> 14.0 = termFreq=14.0
> 3.710244 = idf(docFreq=424, maxDocs=6389)
> 9.765625E-4 = fieldNorm(doc=380)
> 8.445298E-5 = weight(h1:florida in 380) [ClassicSimilarity], result of:
> 8.445298E-5 = score(doc=380,freq=1.0), product of:
> 8.387785E-4 = queryWeight, product of:
> 1000.7 = boost
> 7.3644376 = idf(docFreq=10, maxDocs=6389)
> 1.1381613E-7 = queryNorm
> 0.10068567 = fieldWeight in 380, product of:
> 1.0 = tf(freq=1.0), with freq of:
> 1.0 = termFreq=1.0
> 7.3644376 = idf(docFreq=10, maxDocs=6389)
> 0.013671875 = fieldNorm(doc=380)
> 0.02968075 = weight(metatag.description:florida in 380)
> [ClassicSimilarity], result of:
> 0.02968075 = score(doc=380,freq=1.0), product of:
> 0.3796503 = queryWeight, product of:
> 500000.1 = boost
> 6.6712904 = idf(docFreq=21, maxDocs=6389)
> 1.1381613E-7 = queryNorm
> 0.07817918 = fieldWeight in 380, product of:
> 1.0 = tf(freq=1.0), with freq of:
> 1.0 = termFreq=1.0
> 6.6712904 = idf(docFreq=21, maxDocs=6389)
> 0.01171875 = fieldNorm(doc=380)
> ',
>
> Thanks for any help in understanding and tuning Solr to behave correctly.
>
> Regards
> Megha
