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 "Henrique O. Santos" <he...@gmail.com> on 2015/08/06 21:39:42 UTC

SolrJ update

Hello all,

I am using SolrJ to do a index update on one of my collections. This 
collection has a uniqueKey id field:

   <fields>
     <field name="id" type="string" indexed="true" stored="true"/>
     <field name="_version_" type="long" indexed="true" stored="true"/>
     <field name="name" type="string" indexed="true" stored="true"/>
   </fields>
   <uniqueKey>id</uniqueKey>

This field is configured to be auto generated on solrconfig.xml like this:

<updateRequestProcessorChain>
     <processor class="solr.UUIDUpdateProcessorFactory">
       <str name="fieldName">id</str>
     </processor>
     <processor class="solr.LogUpdateProcessorFactory" />
     <processor class="solr.RunUpdateProcessorFactory" />
</updateRequestProcessorChain>

On my Java code, I just add the name field to my document and then 
proceed with the add:

doc.addField("name", this.name);
solrClient.add(doc);
solrClient.commit();

Everything works, the document gets indexed. What I really need is to 
know right away in the code the id that was generated for that single 
document. I have tried looking into the UpdateReponse but no luck.

Is there any easy way to do that?

Thank you in advance,
Henrique.

Re: SolrJ update

Posted by "Henrique O. Santos" <he...@gmail.com>.
Hi Andrea,

Thanks for the explanation on that. I ended up doing that, getting and 
setting an UUID on the Java code.

On 08/08/2015 03:19 AM, Andrea Gazzarini wrote:
> Hi Henrique,
> I don't believe there's an easy way to do that.
>
> As you noticed, the SolrInputDocument is not an I/O param, that is, it is
> not sent back once data has been indexed and this is good, because here
> you're sending just one document, but imagine what could happen if you do a
> bulk loading...the response would be very very huge!
>
> Although I could imagine some workaround (with a custom
> UpdateRequestProcessor and a custom ResponseWriter), the point is that (see
> above) I believe it would end in a bad design:
>
>     - if you send one document at time this is *often* considered a bad
>     practice;
>     - if you send a lot of data the corresponding response would be huge, it
>     will contains a lot of new created identifiers....and BTW how do you match
>     them with your input documents? Sequentially? mmmm....in this way you won't
>     be able to use any *asynch* client
>
> Personally, if that is ok for your context, I'd completely avoid the
> problem moving the logic on the client side. I mean, create a UUID field on
> Solrj and add that ID to the outcoming document.
>
> Best,
> Andrea
>
>
>
> 2015-08-06 21:39 GMT+02:00 Henrique O. Santos <he...@gmail.com>:
>
>> Hello all,
>>
>> I am using SolrJ to do a index update on one of my collections. This
>> collection has a uniqueKey id field:
>>
>>    <fields>
>>      <field name="id" type="string" indexed="true" stored="true"/>
>>      <field name="_version_" type="long" indexed="true" stored="true"/>
>>      <field name="name" type="string" indexed="true" stored="true"/>
>>    </fields>
>>    <uniqueKey>id</uniqueKey>
>>
>> This field is configured to be auto generated on solrconfig.xml like this:
>>
>> <updateRequestProcessorChain>
>>      <processor class="solr.UUIDUpdateProcessorFactory">
>>        <str name="fieldName">id</str>
>>      </processor>
>>      <processor class="solr.LogUpdateProcessorFactory" />
>>      <processor class="solr.RunUpdateProcessorFactory" />
>> </updateRequestProcessorChain>
>>
>> On my Java code, I just add the name field to my document and then proceed
>> with the add:
>>
>> doc.addField("name", this.name);
>> solrClient.add(doc);
>> solrClient.commit();
>>
>> Everything works, the document gets indexed. What I really need is to know
>> right away in the code the id that was generated for that single document.
>> I have tried looking into the UpdateReponse but no luck.
>>
>> Is there any easy way to do that?
>>
>> Thank you in advance,
>> Henrique.
>>


Re: SolrJ update

Posted by Andrea Gazzarini <a....@gmail.com>.
Hi Henrique,
I don't believe there's an easy way to do that.

As you noticed, the SolrInputDocument is not an I/O param, that is, it is
not sent back once data has been indexed and this is good, because here
you're sending just one document, but imagine what could happen if you do a
bulk loading...the response would be very very huge!

Although I could imagine some workaround (with a custom
UpdateRequestProcessor and a custom ResponseWriter), the point is that (see
above) I believe it would end in a bad design:

   - if you send one document at time this is *often* considered a bad
   practice;
   - if you send a lot of data the corresponding response would be huge, it
   will contains a lot of new created identifiers....and BTW how do you match
   them with your input documents? Sequentially? mmmm....in this way you won't
   be able to use any *asynch* client

Personally, if that is ok for your context, I'd completely avoid the
problem moving the logic on the client side. I mean, create a UUID field on
Solrj and add that ID to the outcoming document.

Best,
Andrea



2015-08-06 21:39 GMT+02:00 Henrique O. Santos <he...@gmail.com>:

> Hello all,
>
> I am using SolrJ to do a index update on one of my collections. This
> collection has a uniqueKey id field:
>
>   <fields>
>     <field name="id" type="string" indexed="true" stored="true"/>
>     <field name="_version_" type="long" indexed="true" stored="true"/>
>     <field name="name" type="string" indexed="true" stored="true"/>
>   </fields>
>   <uniqueKey>id</uniqueKey>
>
> This field is configured to be auto generated on solrconfig.xml like this:
>
> <updateRequestProcessorChain>
>     <processor class="solr.UUIDUpdateProcessorFactory">
>       <str name="fieldName">id</str>
>     </processor>
>     <processor class="solr.LogUpdateProcessorFactory" />
>     <processor class="solr.RunUpdateProcessorFactory" />
> </updateRequestProcessorChain>
>
> On my Java code, I just add the name field to my document and then proceed
> with the add:
>
> doc.addField("name", this.name);
> solrClient.add(doc);
> solrClient.commit();
>
> Everything works, the document gets indexed. What I really need is to know
> right away in the code the id that was generated for that single document.
> I have tried looking into the UpdateReponse but no luck.
>
> Is there any easy way to do that?
>
> Thank you in advance,
> Henrique.
>