You are viewing a plain text version of this content. The canonical link for it is here.
Posted to solr-user@lucene.apache.org by Abhijit Pawar <ab...@gmail.com> on 2017/08/11 17:19:20 UTC
How to boost entire json document on top of results
Hi,
I am working on a scenario trying to boost certain documents while indexing
over other documents.
I am using a DataImportHandler to import JSON documents from a MongoDB
datasource.
Here's my data-source-config file :
<dataConfig>
<dataSource name="mongod" type="JdbcDataSource"
driver="com.mongodb.jdbc.MongoDriver" url="mongodb://<<HOST>>:27017/CORE"/>
<script><![CDATA[
function BoostDoc(row) {
if(row.get('boost') == '10') {
row.put('$docBoost', 10);
}
if(row.get('boost') == '20') {
row.put('$docBoost', 20);
}
if(row.get('boost') == '30') {
row.put('$docBoost', 30);
}
return row;
}
]]></script>
<document name="products">
<entity name="products"
dataSource="mongod"
transformer="ProdsCatsFieldTransformer,script:BoostDoc,TemplateTransformer"
onError="continue"
pk="uuid"
query="SELECT idStr,orgidStr,name,boost,code,description....
FROM products"
deltaImportQuery="SELECT idStr,orgidStr,name,boost,code,description....
FROM products"...
deltaQuery="SELECT idStr FROM products WHERE orgidStr ....
>
<field column="idStr" name="prod_idStr"/>
<field column="uuid" name="uuid" template="prod-${products.idStr}"/>
<field column="orgidStr" name="prod_org"/>
<field column="name" name="prod_name"/>
<field column="boost" name="prod_boost"/>
I have added omiNorms to unique identifier - uuid.
With this configuration I am unable to see any change in the results
returned while search.Basically I would like to have documents with boost
value 30 to be shown above the documents with value 10 or 20.
Is there change I need to make in order to apply boost values using
$docBoost with ScriptTransformer ?
Also any other alternative approach that I may use for achieving the
boosted documents to show up in the results list.
Regards,
Abhijit
Re: How to boost entire json document on top of results
Posted by Abhijit Pawar <ab...@gmail.com>.
Yes...That's true for now I can sort based on that field but was wondering
how index time boost really works at the document level for JSON documents
indexed using an DataImportHandler.
Thanks Eric!
Regards,
Abhijit
On Fri, Aug 11, 2017 at 1:50 PM, Erick Erickson <er...@gmail.com>
wrote:
> Well, you have to _do_ something with the field. And you don't
> particularly want to do document boosting at index time in the first
> place as this has been removed recently IIRC. Note this is different
> than just putting a _value_ in some field you _use_ for boosting......
>
> Anyway, you state:
> "I would like to have documents with boost value 30 to be shown above
> the documents with value 10 or 20"
>
> If it's acceptable that all docs with a value of 30 appear before all
> docs with a value of 20 which appear before all docs with a value of
> 10, then this is just a simple sort
>
> sort=boost desc,score desc
>
> If you want to do something fancier, see Function Queries in the ref
> guide. Or use edismax query parsing and add a boost query
>
> Best,
> Erick
>
>
>
> On Fri, Aug 11, 2017 at 10:19 AM, Abhijit Pawar
> <ab...@gmail.com> wrote:
> > Hi,
> >
> > I am working on a scenario trying to boost certain documents while
> indexing
> > over other documents.
> >
> > I am using a DataImportHandler to import JSON documents from a MongoDB
> > datasource.
> >
> > Here's my data-source-config file :
> >
> > <dataConfig>
> > <dataSource name="mongod" type="JdbcDataSource"
> > driver="com.mongodb.jdbc.MongoDriver" url="mongodb://<<HOST>>:27017/
> CORE"/>
> > <script><![CDATA[
> > function BoostDoc(row) {
> > if(row.get('boost') == '10') {
> > row.put('$docBoost', 10);
> > }
> > if(row.get('boost') == '20') {
> > row.put('$docBoost', 20);
> > }
> > if(row.get('boost') == '30') {
> > row.put('$docBoost', 30);
> > }
> > return row;
> > }
> > ]]></script>
> > <document name="products">
> > <entity name="products"
> > dataSource="mongod"
> > transformer="ProdsCatsFieldTransformer,script:BoostDoc,Templ
> ateTransformer"
> > onError="continue"
> > pk="uuid"
> > query="SELECT idStr,orgidStr,name,boost,code,description....
> > FROM products"
> > deltaImportQuery="SELECT idStr,orgidStr,name,boost,code,description....
> > FROM products"...
> > deltaQuery="SELECT idStr FROM products WHERE orgidStr ....
> >
> >>
> > <field column="idStr" name="prod_idStr"/>
> > <field column="uuid" name="uuid" template="prod-${products.idStr}"/>
> > <field column="orgidStr" name="prod_org"/>
> > <field column="name" name="prod_name"/>
> > <field column="boost" name="prod_boost"/>
> >
> > I have added omiNorms to unique identifier - uuid.
> >
> >
> > With this configuration I am unable to see any change in the results
> > returned while search.Basically I would like to have documents with boost
> > value 30 to be shown above the documents with value 10 or 20.
> >
> > Is there change I need to make in order to apply boost values using
> > $docBoost with ScriptTransformer ?
> >
> > Also any other alternative approach that I may use for achieving the
> > boosted documents to show up in the results list.
> >
> >
> >
> > Regards,
> >
> > Abhijit
>
Re: How to boost entire json document on top of results
Posted by Erick Erickson <er...@gmail.com>.
Well, you have to _do_ something with the field. And you don't
particularly want to do document boosting at index time in the first
place as this has been removed recently IIRC. Note this is different
than just putting a _value_ in some field you _use_ for boosting......
Anyway, you state:
"I would like to have documents with boost value 30 to be shown above
the documents with value 10 or 20"
If it's acceptable that all docs with a value of 30 appear before all
docs with a value of 20 which appear before all docs with a value of
10, then this is just a simple sort
sort=boost desc,score desc
If you want to do something fancier, see Function Queries in the ref
guide. Or use edismax query parsing and add a boost query
Best,
Erick
On Fri, Aug 11, 2017 at 10:19 AM, Abhijit Pawar
<ab...@gmail.com> wrote:
> Hi,
>
> I am working on a scenario trying to boost certain documents while indexing
> over other documents.
>
> I am using a DataImportHandler to import JSON documents from a MongoDB
> datasource.
>
> Here's my data-source-config file :
>
> <dataConfig>
> <dataSource name="mongod" type="JdbcDataSource"
> driver="com.mongodb.jdbc.MongoDriver" url="mongodb://<<HOST>>:27017/CORE"/>
> <script><![CDATA[
> function BoostDoc(row) {
> if(row.get('boost') == '10') {
> row.put('$docBoost', 10);
> }
> if(row.get('boost') == '20') {
> row.put('$docBoost', 20);
> }
> if(row.get('boost') == '30') {
> row.put('$docBoost', 30);
> }
> return row;
> }
> ]]></script>
> <document name="products">
> <entity name="products"
> dataSource="mongod"
> transformer="ProdsCatsFieldTransformer,script:BoostDoc,TemplateTransformer"
> onError="continue"
> pk="uuid"
> query="SELECT idStr,orgidStr,name,boost,code,description....
> FROM products"
> deltaImportQuery="SELECT idStr,orgidStr,name,boost,code,description....
> FROM products"...
> deltaQuery="SELECT idStr FROM products WHERE orgidStr ....
>
>>
> <field column="idStr" name="prod_idStr"/>
> <field column="uuid" name="uuid" template="prod-${products.idStr}"/>
> <field column="orgidStr" name="prod_org"/>
> <field column="name" name="prod_name"/>
> <field column="boost" name="prod_boost"/>
>
> I have added omiNorms to unique identifier - uuid.
>
>
> With this configuration I am unable to see any change in the results
> returned while search.Basically I would like to have documents with boost
> value 30 to be shown above the documents with value 10 or 20.
>
> Is there change I need to make in order to apply boost values using
> $docBoost with ScriptTransformer ?
>
> Also any other alternative approach that I may use for achieving the
> boosted documents to show up in the results list.
>
>
>
> Regards,
>
> Abhijit