Faceting with the option of multi-select

Facets are extracted from the search result. Once a customer selects an option from the facet, we create a filter query that is typically an AND logic. For example, on searching for a particular item, say tommy hilfiger, we would be getting results that will have facets for size and color. It was previously assumed that the customer would select a single option from both the facets. Say the selections are medium for size and green for color. The filter query would be:

fq=clothes_size:medium&fq=clothes_color:green

This will be appended to our search query:

q=tommy%20hilfiger&qf=text%20cat^2%20name^2%20brand^2%20clothes_type^2%20clothes_color^2%20clothes_occassion^2&pf=text%20cat^3%20name^3%20brand^3%20clothes_type^3%20clothes_color^3%20clothes_occassion^3&fl=name,clothes_size,clothes_color,score&defType=edismax&facet=true&facet.mincount=1&facet.field=clothes_gender&facet.field=clothes_type&facet.field=clothes_size&facet.field=clothes_color&facet.field=brand&facet.field=mobile_os&facet.field=mobile_screen_size&facet.field=laptop_processor&facet.field=laptop_memory&facet.field=laptop_hard_disk&fq=clothes_size:medium&fq=clothes_color:green 

What happens if the customer intends to select multiple options in facets? In this case, suppose the customer is interested in both medium and large sizes and also in green and red colors. To handle this, we will have to use OR between multiple options in our filter query for a particular field. The filter query in this case will be:

fq=clothes_size:medium clothes_size:large&fq=clothes_color:red clothes_color:green

The complete query will be:

q=tommy%20hilfiger&qf=text%20cat^2%20name^2%20brand^2%20clothes_type^2%20clothes_color^2%20clothes_occassion^2&pf=text%20cat^3%20name^3%20brand^3%20clothes_type^3%20clothes_color^3%20clothes_occassion^3&fl=name,clothes_size,clothes_color,score&defType=edismax&facet=true&facet.mincount=1&facet.field=clothes_gender&facet.field=clothes_type&facet.field=clothes_size&facet.field=clothes_color&facet.field=brand&facet.field=mobile_os&facet.field=mobile_screen_size&facet.field=laptop_processor&facet.field=laptop_memory&facet.field=laptop_hard_disk&fq=clothes_size:mediumclothes_size:large&fq=clothes_color:redclothes_color:green 

The filter query is interpreted as follows:

Filter query for the multi-select option

The results would be as follows:

Multi-select search result

Another problem that plagues faceting on an e-commerce site is known as disappearing facets. To show only facets that are relevant to the search result, we set the facet.mincount parameter to 1. This ensures that only facets that contain certain values are shown. What if further filter queries reduce the facet count to 0 and make it disappear from the facets?

Let us take an example to understand the problem.

A search for tommy hilfiger gives us facets in size and color:

Facets for "tommy hilfiger" search

When we apply a filter query for size=small, we intend to have an output specifying that there are no sweaters in red matching the size we have selected:

size: small -> 1, medium -> 0, large -> 0 color: red -> 0, green -> 1

However, since the facet.mincount=1 parameter, the output is:

size: small -> 1 color: green -> 1

Disappearing facets

The way to handle this is to tag filter queries and exclude them from facets. The filter query can be tagged as follows:

fq={!tag=size_fq}clothes_size:small

The tagged filter query can be excluded from the facet using the following syntax:

facet.field={!ex=size_fq}clothes_size

In order to handle the disappearing facet scenario, we have created multiple facets on the same field—one facet bearing the exact count from the result set taking into consideration the filter queries and another facet that excludes all filter queries.

We have modified the preceding query to handle this scenario, and simplified the query to focus exactly on the modifications that we have implemented:

q=tommy%20hilfiger&qf=text%20cat^2%20name^2%20brand^2%20clothes_type^2%20clothes_color^2%20clothes_occassion^2&pf=text%20cat^3%20name^3%20brand^3%20clothes_type^3%20clothes_color^3%20clothes_occassion^3&fl=name,clothes_size,clothes_color,score&defType=edismax&facet=true&facet.mincount=1&facet.field=clothes_size&facet.field=clothes_color&facet.field={!ex=size_fq,color_fq key=all_size_facets}clothes_size&facet.field={!ex=size_fq,color_fq key=all_color_facets}clothes_color&fq={!tag=size_fq}clothes_size:small 

Here, we have created multiple facets on both clothes_size and clothes_color fields, one in which filter queries are excluded and another in which they are not. Also, the filter queries are tagged as size_fq. The facets in which filter queries are excluded are tagged as all_size_facets and all_color_facets. The following is the output from the query:

Both missing facets and all results facets seen here

Our program has to take care of merging both the facets for size and color and create the appropriate output. For example, the output can contain the facet counts from the all_clothes_color tag. Except for green, all others could be disabled (instead of displaying 0).