Elasticsearch: Find substring match, Exact match and only if present match - elasticsearch

Question

I want to perform such kind of query such that the query shows output if and only if all the words in the query are present in the given string as a string or query
For example -
let text = "garbage can"
so if I query
"garb"
it should return "garbage can"
if I query
"garbage ca"
it should return "garbage can"
but if I query
"garbage b"
it should not return anything
I tried using substring and also match but they both did not quite did the job for me.

Answer 1

You may use an Edge N-gram tokenizer to index your data.
You can also have custom token_chars in the latest 7.8 version!
Have a look at the documentation for more details: https://www.elastic.co/guide/en/elasticsearch/reference/current/analysis-edgengram-tokenizer.html

Answer 2

I guess you want to do a prefix query. please try the following prefix query:
GET /test_index/_search
{
"query": {
"prefix": {
"my_keyword": {
"value": "garbage b"
}
}
}
}
However, this kind of prefix query's performance is not good.
You could try the following query by using customized prefix analyser.
First, create a new index:
PUT /test_index
{
"settings": {
"index": {
"number_of_shards": "1",
"analysis": {
"filter": {
"autocomplete_filter": {
"type": "edge_ngram",
"min_gram": "1",
"max_gram": "20"
}
},
"analyzer": {
"autocomplete": {
"filter": [
"lowercase",
"autocomplete_filter"
],
"type": "custom",
"tokenizer": "keyword"
}
}
},
"number_of_replicas": "1"
}
},
"mappings": {
"properties": {
"my_text": {
"analyzer": "autocomplete",
"type": "text"
},
"my_keyword": {
"type": "keyword"
}
}
}
}
Second, insert data into this index:
PUT /test_index/_doc/1
{
"my_text": "garbage can",
"my_keyword": "garbage can"
}
Query with "garbage c"
GET /test_index/_search
{
"query": {
"term": {
"my_text": "garbage c"
}
}
}
{
"took" : 0,
"timed_out" : false,
"_shards" : {
"total" : 1,
"successful" : 1,
"skipped" : 0,
"failed" : 0
},
"hits" : {
"total" : {
"value" : 1,
"relation" : "eq"
},
"max_score" : 0.45802015,
"hits" : [
{
"_index" : "test_index",
"_type" : "_doc",
"_id" : "1",
"_score" : 0.45802015,
"_source" : {
"my_text" : "garbage can",
"my_keyword" : "garbage can"
}
}
]
}
}
Query with "garbage b"
GET /test_index/_search
{
"query": {
"term": {
"my_text": "garbage b"
}
}
}
{
"took" : 0,
"timed_out" : false,
"_shards" : {
"total" : 1,
"successful" : 1,
"skipped" : 0,
"failed" : 0
},
"hits" : {
"total" : {
"value" : 0,
"relation" : "eq"
},
"max_score" : null,
"hits" : [ ]
}
}
If you don't want to do a prefix query, you could try the following wildcard query. Please remember the performance is bad and you could also try to use customeized analyser to optimize it.
GET /test_index/_search
{
"query": {
"wildcard": {
"my_keyword": {
"value": "*garbage c*"
}
}
}
}
New Edit Part
I'm not sure if I got want you really want this time....
Anyway, please try to use the following _mapping and queries:
1. Create Index
PUT /test_index
{
"settings": {
"index": {
"max_ngram_diff": 50,
"number_of_shards": "1",
"analysis": {
"filter": {
"autocomplete_filter": {
"type": "ngram",
"min_gram": 1,
"max_gram": 51,
"token_chars": [
"letter",
"digit"
]
}
},
"analyzer": {
"autocomplete": {
"filter": [
"lowercase",
"autocomplete_filter"
],
"type": "custom",
"tokenizer": "keyword"
}
}
},
"number_of_replicas": "1"
}
},
"mappings": {
"properties": {
"my_text": {
"analyzer": "autocomplete",
"type": "text"
},
"my_keyword": {
"type": "keyword"
}
}
}
}
2. Insert some smaple data
PUT /test_index/_doc/1
{
"my_text": "test garbage can",
"my_keyword": "test garbage can"
}
PUT /test_index/_doc/2
{
"my_text": "garbage",
"my_keyword": "garbage"
}
3. Query
GET /test_index/_search
{
"query": {
"term": {
"my_text": "bage c"
}
}
}
Please Note:
This index only support string which max length is 50. Otherwise you need to modified max_ngram_diff, min_gram and max_gram
It need a lot of mem to build the reversing index