Search a Elastic query by Length of string - node.js

Question

{
"query": {
"bool": {
"must": {
"bool": {
"must": [
{
"multi_match": {
"query": "FGTHSF-2124-6",
"type": "phrase_prefix",
"fields": [
"contact.name"
]
}
},
{
"terms": {
"contact.id": [
"sdfwerwe",
"6789",
"4567",
"12345"
]
}
}
],
"should": [],
"must_not": []
}
},
"filter": {
"bool": {
"must": {
"bool": {
"must": [],
"should": [],
"must_not": []
}
}
}
}
}
}
}
I have this query if I search it I get no results because the lasst character is single digit but if I give 3 digit it searches proper record...Is there any by default size for phrase prefix query and if it has then how to change it.
I tried giving default operator, prefix_length, max expansions etc.

Answer 1

contact.name has search_analyzer as "standard" and index analyzer as "autocomplete"(settings not available in question). Your issue is because field is converted to different tokens while indexing and searching so they are not matching.
Usually, the same analyzer should be applied at index time and at
search time, to ensure that the terms in the query are in the same
format as the terms in the inverted index.
Sometimes, though, it can make sense to use a different analyzer at
search time, such as when using the edge_ngram tokenizer for
autocomplete.
ex. with below settings
"settings": {
"analysis": {
"filter": {
"autocomplete_filter": {
"type": "edge_ngram",
"min_gram": 3,
"max_gram": 3
}
},
"analyzer": {
"autocomplete": {
"type": "custom",
"tokenizer": "standard",
"filter": [
"lowercase",
"autocomplete_filter"
]
}
}
}
},
"mappings": {
.....
}
Following tokens will be stored in index
GET index79/_analyze
{
"text": "FGTHSF-2124-645",
"analyzer": "autocomplete"
}
{
"tokens" : [
{
"token" : "fgt",
"start_offset" : 0,
"end_offset" : 6,
"type" : "<ALPHANUM>",
"position" : 0
},
{
"token" : "212",
"start_offset" : 7,
"end_offset" : 11,
"type" : "<NUM>",
"position" : 1
},
{
"token" : "645",
"start_offset" : 12,
"end_offset" : 15,
"type" : "<NUM>",
"position" : 2
}
]
}
Now when you are searching for "query": "HSF-2124-6", it will not return any document(["HSF","2124","6"]) as 6 is not present in any token. To return document you need to change "min_gram" to 1 to generate tokens of size 1 , 3 and 3(645=> 6 64 645).
If you are not using edge gram then please add autocomplete definition
Edit 1:
If you will see in your snapshot min gram size is 5 and max gram 20. Minimum token generated is of size 5 ex "Abhijeet" tokens generated for these will be "Abhij","Abhije","Abhijee","Abhjieet". So any text less that size 5 will not match ex "Abhi". In your case take text after splitting on hyphen("-") so 6 is not matching.
You need to update your index settings and make min-gram:1.
Steps.
1. POST /index80/_close
Below is just an example in your settings copy entire analysis part , update mingram and so a PUT request
PUT index80/_settings
{
"analysis": {
"analyzer": {
"my_analyzer": {
"tokenizer": "my_tokenizer"
}
},
"tokenizer": {
"my_tokenizer": {
"type": "edge_ngram",
"min_gram": 1,
"max_gram": 10,
"token_chars": [
"letter",
"digit"
]
}
}
}
}
3. POST /index80/_open
Check here on how to update settings for further info.
Note: Reducing min size will cause creation of additional tokens and will increase size of your index. Choice of min and max gram should be based on your query text size