You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@solr.apache.org by Nishanth Nayakanti <Ni...@tadigital.com.INVALID> on 2021/08/10 13:05:45 UTC

Field value feature returning only 0s - LTR feature engineering

Hello everyone,

We are using Solr 8.5 and facing the following issue.

We want to use number of clicks of a documents/page as a feature in LTR model. We are using in-place updates to set/upload the number of clicks for each document.  When I try to extract features from Solr, all the documents/rows have values as 0 for clicks feature.

Sample feature
{
      "name":"pmetrics_its_nid-ga_metric1-sum",
      "class":"org.apache.solr.ltr.feature.FieldValueFeature",
      "params":{"field":"pmetrics_its_nid-ga_metric1-sum"},
      "store":"_DEFAULT_"
}

Solr Config

<dynamicField name="pmetrics_*" type="pfloat" indexed="false" stored="false" docValues="true"/>


Am I missing something here ?

Thanks for your answers
Nishanth.
Disclaimer: This message and any attachments are solely intended for the addressee(s). It may also be TA DIGITAL confidential, privileged and / or subject to copyright. Access to this email by anyone else is unauthorized. If you are not the intended recipient, any disclosure, copying, distribution or any action taken or omitted to be taken in reliance on it, is prohibited and may be unlawful. If you have received this in error, please notify the sender immediately by return e-mail and delete it from your computer. While all care has been taken, TA DIGITAL management disclaims all liabilities for loss or damages to person(s) or properties arising from misuse of any information provided or the message being infected by computer virus or other contamination.

Re: Field value feature returning only 0s - LTR feature engineering

Posted by Alessandro Benedetti <a....@sease.io>.
Looking at the code of the FieldValueFeature in the LTR integration it
seems to be fully compatible with "DocValues" only fields :
org/apache/solr/ltr/feature/FieldValueFeature.java:128

It would require some additional investigation to understand why you get
only zeroes (have you tried with a different fieldType for example?)
--------------------------
Alessandro Benedetti
Apache Lucene/Solr Committer
Director, R&D Software Engineer, Search Consultant

www.sease.io


On Tue, 10 Aug 2021 at 16:25, Nishanth Nayakanti
<Ni...@tadigital.com.invalid> wrote:

> Hi Dave,
>
> Thanks for responding
>
> I think that config setting is a pre-requisite for inplace updates as
> defined here
> https://solr.apache.org/guide/6_6/updating-parts-of-documents.html#UpdatingPartsofDocuments-Example.1
>
> Here is from Solr documentation.
>
> An atomic update operation is performed using this approach only when the
> fields to be updated meet these three conditions:
>
> are non-indexed (indexed="false"), non-stored (stored="false"), single
> valued (multiValued="false") numeric docValues (docValues="true") fields;
>
> the _version_ field is also a non-indexed, non-stored single valued
> docValues field; and,
>
> copy targets of updated fields, if any, are also non-indexed, non-stored
> single valued numeric docValues fields.
>
> Thanks
> Nishanth
>
> -----Original Message-----
> From: Dave <ha...@gmail.com>
> Sent: Tuesday, August 10, 2021 6:47 PM
> To: users@solr.apache.org
> Cc: Madhuluck Kumar <ma...@tadigital.com>
> Subject: Re: Field value feature returning only 0s - LTR feature
> engineering
>
> ***Caution: External Email***
>
> This email originated from an external sender. Do not click on links or
> open attachments unless you were expecting this message from this sender.
>
> I’m confused. You don’t store it, nor index it?
>
> > On Aug 10, 2021, at 9:06 AM, Nishanth Nayakanti <
> Nishanth.n@tadigital.com.invalid> wrote:
> >
> > Hello everyone,
> >
> > We are using Solr 8.5 and facing the following issue.
> >
> > We want to use number of clicks of a documents/page as a feature in LTR
> model. We are using in-place updates to set/upload the number of clicks for
> each document.  When I try to extract features from Solr, all the
> documents/rows have values as 0 for clicks feature.
> >
> > Sample feature
> > {
> >      "name":"pmetrics_its_nid-ga_metric1-sum",
> >      "class":"org.apache.solr.ltr.feature.FieldValueFeature",
> >      "params":{"field":"pmetrics_its_nid-ga_metric1-sum"},
> >      "store":"_DEFAULT_"
> > }
> >
> > Solr Config
> >
> > <dynamicField name="pmetrics_*" type="pfloat" indexed="false"
> stored="false" docValues="true"/>
> >
> >
> > Am I missing something here ?
> >
> > Thanks for your answers
> > Nishanth.
> > Disclaimer: This message and any attachments are solely intended for the
> addressee(s). It may also be TA DIGITAL confidential, privileged and / or
> subject to copyright. Access to this email by anyone else is unauthorized.
> If you are not the intended recipient, any disclosure, copying,
> distribution or any action taken or omitted to be taken in reliance on it,
> is prohibited and may be unlawful. If you have received this in error,
> please notify the sender immediately by return e-mail and delete it from
> your computer. While all care has been taken, TA DIGITAL management
> disclaims all liabilities for loss or damages to person(s) or properties
> arising from misuse of any information provided or the message being
> infected by computer virus or other contamination.
> Disclaimer: This message and any attachments are solely intended for the
> addressee(s). It may also be TA DIGITAL confidential, privileged and / or
> subject to copyright. Access to this email by anyone else is unauthorized.
> If you are not the intended recipient, any disclosure, copying,
> distribution or any action taken or omitted to be taken in reliance on it,
> is prohibited and may be unlawful. If you have received this in error,
> please notify the sender immediately by return e-mail and delete it from
> your computer. While all care has been taken, TA DIGITAL management
> disclaims all liabilities for loss or damages to person(s) or properties
> arising from misuse of any information provided or the message being
> infected by computer virus or other contamination.
>

RE: Field value feature returning only 0s - LTR feature engineering

Posted by Nishanth Nayakanti <Ni...@tadigital.com.INVALID>.
Hi Dave,

Thanks for responding

I think that config setting is a pre-requisite for inplace updates as defined here https://solr.apache.org/guide/6_6/updating-parts-of-documents.html#UpdatingPartsofDocuments-Example.1

Here is from Solr documentation.

An atomic update operation is performed using this approach only when the fields to be updated meet these three conditions:

are non-indexed (indexed="false"), non-stored (stored="false"), single valued (multiValued="false") numeric docValues (docValues="true") fields;

the _version_ field is also a non-indexed, non-stored single valued docValues field; and,

copy targets of updated fields, if any, are also non-indexed, non-stored single valued numeric docValues fields.

Thanks
Nishanth

-----Original Message-----
From: Dave <ha...@gmail.com>
Sent: Tuesday, August 10, 2021 6:47 PM
To: users@solr.apache.org
Cc: Madhuluck Kumar <ma...@tadigital.com>
Subject: Re: Field value feature returning only 0s - LTR feature engineering

***Caution: External Email***

This email originated from an external sender. Do not click on links or open attachments unless you were expecting this message from this sender.

I’m confused. You don’t store it, nor index it?

> On Aug 10, 2021, at 9:06 AM, Nishanth Nayakanti <Ni...@tadigital.com.invalid> wrote:
>
> Hello everyone,
>
> We are using Solr 8.5 and facing the following issue.
>
> We want to use number of clicks of a documents/page as a feature in LTR model. We are using in-place updates to set/upload the number of clicks for each document.  When I try to extract features from Solr, all the documents/rows have values as 0 for clicks feature.
>
> Sample feature
> {
>      "name":"pmetrics_its_nid-ga_metric1-sum",
>      "class":"org.apache.solr.ltr.feature.FieldValueFeature",
>      "params":{"field":"pmetrics_its_nid-ga_metric1-sum"},
>      "store":"_DEFAULT_"
> }
>
> Solr Config
>
> <dynamicField name="pmetrics_*" type="pfloat" indexed="false" stored="false" docValues="true"/>
>
>
> Am I missing something here ?
>
> Thanks for your answers
> Nishanth.
> Disclaimer: This message and any attachments are solely intended for the addressee(s). It may also be TA DIGITAL confidential, privileged and / or subject to copyright. Access to this email by anyone else is unauthorized. If you are not the intended recipient, any disclosure, copying, distribution or any action taken or omitted to be taken in reliance on it, is prohibited and may be unlawful. If you have received this in error, please notify the sender immediately by return e-mail and delete it from your computer. While all care has been taken, TA DIGITAL management disclaims all liabilities for loss or damages to person(s) or properties arising from misuse of any information provided or the message being infected by computer virus or other contamination.
Disclaimer: This message and any attachments are solely intended for the addressee(s). It may also be TA DIGITAL confidential, privileged and / or subject to copyright. Access to this email by anyone else is unauthorized. If you are not the intended recipient, any disclosure, copying, distribution or any action taken or omitted to be taken in reliance on it, is prohibited and may be unlawful. If you have received this in error, please notify the sender immediately by return e-mail and delete it from your computer. While all care has been taken, TA DIGITAL management disclaims all liabilities for loss or damages to person(s) or properties arising from misuse of any information provided or the message being infected by computer virus or other contamination.

Re: Field value feature returning only 0s - LTR feature engineering

Posted by Dave <ha...@gmail.com>.
I’m confused. You don’t store it, nor index it?

> On Aug 10, 2021, at 9:06 AM, Nishanth Nayakanti <Ni...@tadigital.com.invalid> wrote:
> 
> Hello everyone,
> 
> We are using Solr 8.5 and facing the following issue.
> 
> We want to use number of clicks of a documents/page as a feature in LTR model. We are using in-place updates to set/upload the number of clicks for each document.  When I try to extract features from Solr, all the documents/rows have values as 0 for clicks feature.
> 
> Sample feature
> {
>      "name":"pmetrics_its_nid-ga_metric1-sum",
>      "class":"org.apache.solr.ltr.feature.FieldValueFeature",
>      "params":{"field":"pmetrics_its_nid-ga_metric1-sum"},
>      "store":"_DEFAULT_"
> }
> 
> Solr Config
> 
> <dynamicField name="pmetrics_*" type="pfloat" indexed="false" stored="false" docValues="true"/>
> 
> 
> Am I missing something here ?
> 
> Thanks for your answers
> Nishanth.
> Disclaimer: This message and any attachments are solely intended for the addressee(s). It may also be TA DIGITAL confidential, privileged and / or subject to copyright. Access to this email by anyone else is unauthorized. If you are not the intended recipient, any disclosure, copying, distribution or any action taken or omitted to be taken in reliance on it, is prohibited and may be unlawful. If you have received this in error, please notify the sender immediately by return e-mail and delete it from your computer. While all care has been taken, TA DIGITAL management disclaims all liabilities for loss or damages to person(s) or properties arising from misuse of any information provided or the message being infected by computer virus or other contamination.