Search Facets: Histogram Facet - allow to create buckets with one field name, and aggregate valued on another field

[kimchy]

You can created buckets based on one field values (for example, timestamp), and total/count another field from the same doc.

Here is an example:

{
    query : {
        "match_all" : {}
    },
    facets : {
        "histo1" : {
            "histogram" : {
                "key_field" : "fieldX",
                "value_field" : "fieldY",
                "time_interval" : "5d"
            }
        }
    }
}

[kimchy]

implemented.

[mootpointer]

What if all I want is the bucketed facets, no aggregation?

[kimchy]

can you give an example?

[mootpointer]

So for example I want to know the the count of documents fitting into date buckets.

Eg. I have a document with a created_at field which is a date. I want to know how many have been found each day of the month for the last month, something akin to http://wiki.apache.org/solr/SimpleFacetParameters#Date_Faceting_Parameters

[kimchy]

The check out #228, it gives you the ability to work within the script on a date object, which has simple constructs to get, for example, the day of month as the key. In order to perform it in the last month, you will need to either execute a filter on the facet, or on the query, depends on the scope of your query (if its already filters by "last month filter).

[kimchy]

One more thing, you can also define time_interval which is in elasticsearch time format (for example 2d or 1.5h) as the interval. Since dates are represented as millis in the index, this will work.