Puzzled by search score

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Puzzled by search score

dm_tim
Howdy,
I'm reading a table in a db using the following schema:
 <fields>   
 
   <field name="id" type="string" indexed="true" stored="true" required="true" /> 
   <field name="cid" type="long" indexed="true" stored="true" required="true"/>
   <field name="lang" type="string" indexed="true" stored="true" required="true"/>
   <field name="file_version" type="int" indexed="true" stored="true" required="true"/>
   <field name="search_id" type="long" indexed="true" stored="true" required="true"/>
   <field name="tag" type="text_general" indexed="true" stored="true" required="true"/>
   <field name="created" type="date" indexed="false" stored="true"/>
   <field name="last_modified" type="date" indexed="true" stored="true"/>
   <field name="version" type="long" indexed="true" stored="true"/>
   <field name="_version_" type="long" indexed="true" stored="true" multiValued="false"/>
 </fields>
 
 
 <uniqueKey>id</uniqueKey>

 
 <defaultSearchField>tag</defaultSearchField>

 
 <solrQueryParser defaultOperator="OR"/>
========================================
So make the following query
http://localhost:8080/apache-solr-4.0.0/core0/select?q=tag%3Aclothes~%2Bcid%3A14&sort=score+desc&rows=10&fl=tag+score&wt=json&indent=true

You will notice that I'm doing a search on the tag field against the string "clothes" and the cid field against the long 14 and requesting that the results come back sorted on descending score values. So I'm surprised to see these results:
{
  "responseHeader":{
    "status":0,
    "QTime":1,
    "params":{
      "q":"tag:clothes~+cid:14",
      "sort":"score desc",
      "rows":"10",
      "fl":"tag score",
      "wt":"json",
      "indent":"true"}},
  "response":{"numFound":1835,"start":0,"maxScore":3.9238024,"docs":[
      {
        "tag":"Table Cloth",
        "score":3.9238024},
      {
        "tag":"Clothes",
        "score":3.9134552},
      {
        "tag":"Clothes",
        "score":3.9134552},
      {
        "tag":"Clothes",
        "score":3.9134552},
      {
        "tag":"Clothes",
        "score":3.9134552},
      {
        "tag":"Clothes",
        "score":3.9134552},
      {
        "tag":"Clothes",
        "score":3.9134552},
      {
        "tag":"Boys Clothes",
        "score":3.3968315},
      {
        "tag":"Everyday Clothes",
        "score":3.3968315},
      {
        "tag":"Designer Clothes",
        "score":3.3968315}]
  }}

Why does "Table Cloth" have a higher score than "Clothes" (which is an exact textual match)? I could use some help understanding why I have these results and how to tweak my query so that the results match my expectations.

Regards,

Tim
Reply | Threaded
Open this post in threaded view
|

Re: Puzzled by search score

Otis Gospodnetić
Append &debugQuery=true to your query URL and scroll down to see what is
contributing to the score and how much.

Otis
--
Search Analytics - http://sematext.com/search-analytics/index.html
Performance Monitoring - http://sematext.com/spm/index.html


On Fri, Nov 2, 2012 at 5:20 PM, dm_tim <[hidden email]> wrote:

> Howdy,
> I'm reading a table in a db using the following schema:
>  <fields>
>
>    <field name="id" type="string" indexed="true" stored="true"
> required="true" />
>    <field name="cid" type="long" indexed="true" stored="true"
> required="true"/>
>    <field name="lang" type="string" indexed="true" stored="true"
> required="true"/>
>    <field name="file_version" type="int" indexed="true" stored="true"
> required="true"/>
>    <field name="search_id" type="long" indexed="true" stored="true"
> required="true"/>
>    <field name="tag" type="text_general" indexed="true" stored="true"
> required="true"/>
>    <field name="created" type="date" indexed="false" stored="true"/>
>    <field name="last_modified" type="date" indexed="true" stored="true"/>
>    <field name="version" type="long" indexed="true" stored="true"/>
>    <field name="_version_" type="long" indexed="true" stored="true"
> multiValued="false"/>
>  </fields>
>
>
>  <uniqueKey>id</uniqueKey>
>
>
>  <defaultSearchField>tag</defaultSearchField>
>
>
>  <solrQueryParser defaultOperator="OR"/>
> ========================================
> So make the following query
>
> http://localhost:8080/apache-solr-4.0.0/core0/select?q=tag%3Aclothes~%2Bcid%3A14&sort=score+desc&rows=10&fl=tag+score&wt=json&indent=true
>
> You will notice that I'm doing a search on the tag field against the string
> "clothes" and the cid field against the long 14 and requesting that the
> results come back sorted on descending score values. So I'm surprised to
> see
> these results:
> {
>   "responseHeader":{
>     "status":0,
>     "QTime":1,
>     "params":{
>       "q":"tag:clothes~+cid:14",
>       "sort":"score desc",
>       "rows":"10",
>       "fl":"tag score",
>       "wt":"json",
>       "indent":"true"}},
>   "response":{"numFound":1835,"start":0,"maxScore":3.9238024,"docs":[
>       {
>         "tag":"Table Cloth",
>         "score":3.9238024},
>       {
>         "tag":"Clothes",
>         "score":3.9134552},
>       {
>         "tag":"Clothes",
>         "score":3.9134552},
>       {
>         "tag":"Clothes",
>         "score":3.9134552},
>       {
>         "tag":"Clothes",
>         "score":3.9134552},
>       {
>         "tag":"Clothes",
>         "score":3.9134552},
>       {
>         "tag":"Clothes",
>         "score":3.9134552},
>       {
>         "tag":"Boys Clothes",
>         "score":3.3968315},
>       {
>         "tag":"Everyday Clothes",
>         "score":3.3968315},
>       {
>         "tag":"Designer Clothes",
>         "score":3.3968315}]
>   }}
>
> Why does "Table Cloth" have a higher score than "Clothes" (which is an
> exact
> textual match)? I could use some help understanding why I have these
> results
> and how to tweak my query so that the results match my expectations.
>
> Regards,
>
> Tim
>
>
>
> --
> View this message in context:
> http://lucene.472066.n3.nabble.com/Puzzled-by-search-score-tp4017904.html
> Sent from the Solr - User mailing list archive at Nabble.com.
>