Cannot train 2 or more features for Solr LTR using LIBLINEAR

Hi, I am using the Solr LTR in Solr 7.4.0, and I am trying to train an example learning model using LIBLINEAR. When I tried to run the code from train_and_upload_demo_model.py , I can only train one feature at a time. If I put more than one features, then I will get the following error

Re: Can the export handler be used with the edismax or dismax query handler

Thanks Mikhail and Erick, I don't need ranks or score. I just need the full set of results. Will the export handler work with a fq that uses edismax? (I'm not at work today, but I can try it out tomorrow.) I compared a simple (not edismax) query and the export handler with cursormark with rows

Re: Search without passing any field

Okay, great. I will start with the available solution . My index is less than yours, so may be it will work. Thanks Renuka Srishti On Mon, Jul 30, 2018 at 12:04 AM, Walter Underwood wrote: > No, it won’t slow down performance, but it might use more disk space for > the index. > > Searching one

Re: Search without passing any field

No, it won’t slow down performance, but it might use more disk space for the index. Searching one field is usually faster than searching multiple fields. If it was a lot slower, the tutorial would have warned about that. How large is your index? There are some very large Solr collections. Ours

Re: Search without passing any field

Thanks I got it. Somehow I have to tell about fields to solr, it can't automatically apply search on all indexed fields. On Sun, Jul 29, 2018 at 11:54 PM, Alexandre Rafalovitch wrote: > You may be doing premature optimization here. > > Remember the copyTarget is store=false/docValues=false, so

Re: Search without passing any field

You may be doing premature optimization here. Remember the copyTarget is store=false/docValues=false, so you are only actually storing unique tokens and document ids/offsets. I would recommend you start from that, do your first schema, figure out what you are not happy about, evolve it, etc.

Re: Search without passing any field

Yeah, I read that, but it will slow down the performance, as copying all the fields and put that data in one field. I think for large index it is not the right way to do. Am I right? Regards, Renuka Srishti On Sun, Jul 29, 2018 at 11:39 PM, Alexandre Rafalovitch wrote: > Did you go through the

Re: Search without passing any field

Did you go through the tutorial in the Reference Guide? it explains a lot of these and has configuration for you to check. See for example (in a middle of tutorial): https://lucene.apache.org/solr/guide/7_4/solr-tutorial.html#create-a-catchall-copy-field Regards, Alex. On 29 July 2018 at

Re: Search without passing any field

Thanks for response, Let me explain with an example. I have following fields : name, description, id, title. These fields have following values: doc1 - name : test description : test and run id : t1 title : abc title doc2 - name : abc description : test and run again id : t2 title : xyz title

Re: Search without passing any field

What do you want Solr to do in that case? Search all the defined fields? Some of them? Pick randomly? This is like asking "can Solr do what I want without telling it what I want?". You can define df or qf _defaults_ in solrconfig.xml for the request handler, that's what those are for. Then the

Re: Search without passing any field

Solr does need to know what field(s) you want to search. And you can configure all of these things in solrconfig.xml, so the user does not have to provide that. Just remember, unless you really know what you are doing, you should not be exposing Solr directly to the user/browser. Think of it more

Search without passing any field

Hi, Is there any possible way to perform search by using "q" or "query" parameter without setting "df" parameter or "qf" parameter. for example: I am passing "q":"xyz" I have to mention the field for which I want to perform query otherwise I will not get results. What is the way if I want to

Re: Parameter Mapping

You are right. Yes please, create a ticket for it. If you can do a patch against the trunk (version 8), it is even better. You can do it directly on Github if doing git checkout is hard. Just make sure to reference the relevant Jira ticket. Thank you, Alex. On 29 July 2018 at 08:39, Renuka

Parameter Mapping

Hi, I was reading the document and I think parameters mapping mentioned here is not right. start, rows are used for query parameters. offset, limit are used for JSON query format. Should I create the ticket for