Is there a way to get related documents if a match occurs after the query? - elasticsearch

Question

I am currently doing a fuzzy name search on some documents. These documents can be related to each other (for example name field of one document may contain the name and another may contain the alias for the same person). I will give these documents the same unique identifier. My question is, can I get the documents with same unique identifier if a match occurs in any of them?
Suppose that there are 4 documents like this.
{
{
"name": "Bob"
"uid": "1"
},
{
"name": "Bilbo"
"uid": "1"
},
{
"name": "Jack"
"uid": "2"
},
{
"name": "Mary"
"uid" : "3"
}
}
When I query name "Bob", I expect to get both documents with "uid" = "1"
{
{
"name": "Bob"
"uid": "1"
},
{
"name": "Bilbo"
"uid": "1"
}
}

Answer 1

Elasticsearch doesn't have concept of JOINS. So documents cannot be fetched by joining on "uid"
1. Using two queries
i. Get documents with name "Bob"
{
"query": {
"term": {
"name.keyword": {
"value": "Bob"
}
}
}
}
ii. Fetch documents using above returned ids.
2. Using terms and bucket selector aggregation
Mapping:
{
"<mapping_name>" : {
"mappings" : {
"properties" : {
"name" : {
"type" : "text",
"fields" : {
"keyword" : {
"type" : "keyword",
"ignore_above" : 256
}
}
},
"uid" : {
"type" : "text",
"fields" : {
"keyword" : {
"type" : "keyword",
"ignore_above" : 256
}
}
}
}
}
}
}
Query:
1. Create a bucket(collection) of uid.
2. Create sub bucket of name which includes only "Bob" so uid 1 will have a bucket of key Bob , uid 2 will be empty
3. Use bucket_selector aggregation to select where count of sub bucket name is greater than equal to 1. This will remove uid 2
4. Use top_hits aggregation to get documents.
{
"size": 0,
"aggs": {
"uid": {
"terms": {
"field": "uid.keyword",
"size": 10
},
"aggs": {
"documents":{
"top_hits": { --> to get documents under parent term
"size": 10
}
},
"name": {
"terms": {
"field": "name.keyword", --> terms need non_analyzed field so keyword
"include":"Bob", --> get terms with name bob
"size": 10
}
},
"my_bucket":{
"bucket_selector": { --> select buckets which have atleast one name
"buckets_path": {"count":"name._bucket_count"},
"script": "if(params.count>=1) return true;"
}
}
}
}
}
}
Result: All docuents with uid 1(same uid as "Bob") are returned
"aggregations" : {
"uid" : {
"doc_count_error_upper_bound" : 0,
"sum_other_doc_count" : 0,
"buckets" : [
{
"key" : "1",
"doc_count" : 2,
"documents" : {
"hits" : {
"total" : {
"value" : 2,
"relation" : "eq"
},
"max_score" : 1.0,
"hits" : [
{
"_index" : "index61",
"_type" : "_doc",
"_id" : "uCP1-nAB_Wo5RvhlZM6k",
"_score" : 1.0,
"_source" : {
"name" : "Bob",
"uid" : "1"
}
},
{
"_index" : "index61",
"_type" : "_doc",
"_id" : "uSP1-nAB_Wo5Rvhlbc4S",
"_score" : 1.0,
"_source" : {
"name" : "Bilbo",
"uid" : "1"
}
}
]
}
},
"name" : {
"doc_count_error_upper_bound" : 0,
"sum_other_doc_count" : 0,
"buckets" : [
{
"key" : "Bob",
"doc_count" : 1
}
]
}
}
]
}
}