Counting comma saperated elements as value in elasticsearch - elasticsearch

Question

Is it possible to count the values ​​separated by commas from the same field with aggregations or any other way?
i.e
JSON
{
"ID":"example.1",
"Sports":{
"1":
{,
"Football Teams":" Real Madrid, Manchester United, Juventus"
"Basket Teams":"Chicago Bulls"
},
"2":
{
"Football Teams":"F.C Barcelona, Milan"
"Basket Teams":"Lakers"
}
}
}
query
GET xxx/_search
{
"query": {
"match_all": {
}
},
"aggs": {
"NAME": {
"value_count": {
"field": "Sports.1."Football Teams.keyword"
}}
}
}
Desired Output
"aggregations" : {
"Count" : {
"doc_count_error_upper_bound" : 0,
"sum_other_doc_count" : 0,
"buckets" : [
{
"key" : "Real Madrid, Manchester United, Juventus",
"doc_count" : 3
}
The objective of the query I am looking for is to be able to determine how many values ​​separated by commas a field has.

Answer 1

You can do it with a value script in a terms sub-aggregation:
GET xxx/_search
{
"aggs": {
"count": {
"terms": {
"field": "team.keyword"
},
"aggs": {
"count": {
"terms": {
"field": "team.keyword",
"script": {
"source": "/,/.split(_value).length",
"lang": "painless"
}
}
}
}
}
}
}
The top-level buckets will be the values of the football team and each top-level bucket will have a sub-bucket with the number of tokens, like this:
{
"key" : "Real Madrid, Manchester United, Juventus", <-- team name
"doc_count" : 1,
"count" : {
"doc_count_error_upper_bound" : 0,
"sum_other_doc_count" : 0,
"buckets" : [
{
"key" : "3", <-- number of teams
"doc_count" : 1
}
]
}
}
In order for this to work, you need to add the following line in elasticsearch.yml and restart your server:
script.painless.regex.enabled: true