ElasticSearch - Exact whole words matching for Equal to operations - elasticsearch

Question

I am looking for a way to get the result by matching EXACT WHOLE WORDS from Elasticsearch. This is for "EQ" ("=") operations from UI.
{
"_index": "docs",
"_type": "_doc",
"_id": "1",
"_score": 1.0,
"_source": {
"DocId": 1,
"DocDate": "2020-07-24T10:16:44.0000000Z",
"Conversation": "I just need to know how frequently I should remind you"
}
},
{
"_index": "docs",
"_type": "_doc",
"_id": "2",
"_score": 1.0,
"_source": {
"DocId": 2,
"DocDate": "2020-07-25T10:16:45.0000000Z",
"Conversation": "Building a work culture in your firm"
}
}
in here, when querying with "I just need to know how frequently I should remind you" for Conversation, then only ES should return DocId 1 data.
Even the query is like "I just need to know how frequently I should remind", then it should return empty.
I tried these ES queries, but not able to figure it out.
GET docs/_search
{
"query": {
"bool": {
"must": [
{"match_phrase": {
"Conversation": "just need to know"
}}
]
}
}
}
GET docs/_search
{
"query": {
"query_string": {
"default_field": "Conversation",
"query": "\"just need to know\""
}
}
}
GET docs/_search
{
"query": {
"bool": {
"must": [
{
"match": {
"Conversation": {"query": "just need to know",
"operator": "and"
}
}
}
]
}
}
}

Answer 1

You need to add .keyword to the Conversation field. This uses the keyword analyzer instead of the standard analyzer (notice the ".keyword" after the Conversation field).
When using standard analyzer
GET /_analyze
{
"analyzer" : "standard",
"text" : "I just need to know how frequently I should remind you"
}
The following tokens are generated :
{
"tokens": [
{
"token": "i",
"start_offset": 0,
"end_offset": 1,
"type": "<ALPHANUM>",
"position": 0
},
{
"token": "just",
"start_offset": 2,
"end_offset": 6,
"type": "<ALPHANUM>",
"position": 1
},
{
"token": "need",
"start_offset": 7,
"end_offset": 11,
"type": "<ALPHANUM>",
"position": 2
},
{
"token": "to",
"start_offset": 12,
"end_offset": 14,
"type": "<ALPHANUM>",
"position": 3
},
{
"token": "know",
"start_offset": 15,
"end_offset": 19,
"type": "<ALPHANUM>",
"position": 4
},
{
"token": "how",
"start_offset": 20,
"end_offset": 23,
"type": "<ALPHANUM>",
"position": 5
},
{
"token": "frequently",
"start_offset": 24,
"end_offset": 34,
"type": "<ALPHANUM>",
"position": 6
},
{
"token": "i",
"start_offset": 35,
"end_offset": 36,
"type": "<ALPHANUM>",
"position": 7
},
{
"token": "should",
"start_offset": 37,
"end_offset": 43,
"type": "<ALPHANUM>",
"position": 8
},
{
"token": "remind",
"start_offset": 44,
"end_offset": 50,
"type": "<ALPHANUM>",
"position": 9
},
{
"token": "you",
"start_offset": 51,
"end_offset": 54,
"type": "<ALPHANUM>",
"position": 10
}
]
}
Whereas the keyword analyzer returns the entire input string as a single token.
If you have not defined any explicit mapping, then your modified search query will be :
{
"query": {
"match": {
"Conversation.keyword": "I just need to know how frequently I should remind you"
}
}
}
You can even change your index mapping, in the following way :
{
"mappings": {
"properties": {
"Conversation": {
"type": "keyword"
}
}
}
}