Similar to the nested aggregator, but in reverse. It allows to create buckets and metrics based on the parent docs in the buckets of child documents.

This aggregation only makes sense inside a nested aggregator. Assume a ticket system with where the comments are embedded into the issue document. The reverse_nested aggregator would look like this:

"aggs" : {
   "top_commenters_to_issues" : {
        "nested" : {"path" : "comments"},
        "aggs" : {
            "top_commenters" : {
                "terms" : { "field" : "username" },
                "aggs" : {
                    "top_tags_to_comments" : {
                        "reverse_nested" : {
                            "path" : "comments"
                        },
                        "aggs" : {
                            "top_tags" : { "terms" : { "field" :  "tags"}}
                        }    
                    }
                }
            }
        }
    }
}

This aggregation would return top username buckets and per username bucket the top tags. The important constraint that the nested and reverse_nested add here is that only tag buckets are associated to username buckets that reside in the nested structure.