Support for "multi-fields" in ES transform about luqum HOT 8 CLOSED

Yes it is supported.

Look at ElasticsearchQueryBuilder there are options to specify the field that are not analyzed (that is are of type keywords).

Those parameters can be deduced from your mapping using SchemaAnalyzer

There is a short demonstration in Quick Start

from luqum.

Thanks for your reply!

However I can't get it to work. With the following (partially) mapping for a single field:

"attributes":{
    "type":"nested",
    "properties":{
        "first_name":{
            "type":"text",
            "fields":{
                "raw":{
                    "type":"keyword"
                }
            }
        }
    }
}

As you can see I'm using a text field which is also a keyword. And fail in getting it working with the following search query: attributes.first_name.raw:<search_query>

from luqum.

Hum I don't think the schema analyzer support "fields" right now, nor the ElasticsearchQueryBuilder

This could be a good addition of course. I think it's like having an object_field, but relaxing a bit to enable calling it directly (and of course this as to be separate to be more understandable).

I may look into this, but I can't promise to have it quickly. If you are interested in making a PR it is welcome. I may guide you if you have questions after a first look at the code.

from luqum.

@sefaYavuz, if you use None for object_fields parameter of ElasticsearchQueryBuilder it should work !

from luqum.

@sefaYavuz I just pushed a commit fixing this problem.

You now have a sub_fields parameter on ElasticSearchQueryBuilder.

Could you test if it's ok on your side. If so, I would make a release.

from luqum.

@alexgarel Sorry was busy with other stuff in the meanwhile. I will test it today and will get back to you.

from luqum.

@alexgarel Works as expected! When will you be able to make it into a release?

from luqum.

@sefaYavuz released : https://pypi.org/project/luqum/0.7.5/

from luqum.