Elasticsearch - Index Mapping settings for both exact and partial matching - indexing

Question

I'm new to elasticsearch and am trying to learn how to index using optimal mapping settings to achieve the following.
If I have a document like this
{"name":"Galapagos Islands"}
I want to get this a result for both the following queries
1) Partial matching
{
"query": {
"match": {
"name": "ga"
}
}
}
2) Exact matching
{
"query": {
"term": {
"name": "Galapagos Islands"
}
}
}
With the setting I have currently. I am able to achieve the partial matching part. But exact matching returns no results. Please find below the settings with which I indexed.
{
"mappings": {
"islands": {
"properties": {
"name":{
"type": "string",
"index_analyzer": "autocomplete",
"search_analyzer": "search_ngram"
}
}
}
},
"settings":{
"analysis":{
"analyzer":{
"autocomplete":{
"type":"custom",
"tokenizer":"standard",
"filter":[ "standard", "lowercase", "stop", "kstem", "ngram" ]
},
"search_ngram": {
"type": "custom",
"tokenizer": "keyword",
"filter": "lowercase"
}
},
"filter":{
"ngram":{
"type":"ngram",
"min_gram":2,
"max_gram":15
}
}
}
}
}
What is the correct way to do exact matching and partial matching on a field ?
UPDATE
After recreating the index with settings given below. My mappings look like this
curl -XGET 'localhost:9200/testing/_mappings?pretty'
{
"testing" : {
"mappings" : {
"islands" : {
"properties" : {
"name" : {
"type" : "string",
"index_analyzer" : "autocomplete",
"search_analyzer" : "search_ngram",
"fields" : {
"raw" : {
"type" : "string",
"analyzer" : "my_keyword_lowercase_analyzer"
}
}
}
}
}
}
}
}
My indexing settings are the below
{
"mappings": {
"islands": {
"properties": {
"name":{
"type": "string",
"index_analyzer": "autocomplete",
"search_analyzer": "search_ngram",
"fields": {
"raw": {
"type": "string",
"analyzer": "my_keyword_lowercase_analyzer"
}
}
}
}
}
},
"settings":{
"analysis":{
"analyzer":{
"autocomplete":{
"type":"custom",
"tokenizer":"standard",
"filter":[ "standard", "lowercase", "stop", "kstem", "ngram" ]
},
"search_ngram": {
"type": "custom",
"tokenizer": "keyword",
"filter": "lowercase"
},
"my_keyword_lowercase_analyzer": {
"type": "custom",
"filter": ["lowercase"],
"tokenizer": "keyword"
}
},
"filter":{
"ngram":{
"type":"ngram",
"min_gram":2,
"max_gram":15
}
}
}
}
}
And with all the above, when I query like this
curl -XGET 'localhost:9200/testing/islands/_search?pretty' -d '{"query": {"term": {"name.raw" : "Galapagos Islands"}}}'
{
"took" : 1,
"timed_out" : false,
"_shards" : {
"total" : 5,
"successful" : 5,
"failed" : 0
},
"hits" : {
"total" : 0,
"max_score" : null,
"hits" : [ ]
}
}
And My document is this
curl -XGET 'localhost:9200/testing/islands/1?pretty'
{
"_index" : "testing",
"_type" : "islands",
"_id" : "1",
"_version" : 1,
"found" : true,
"_source":{"name":"Galapagos Islands"}
}

Answer 1

Add a subfield to your name property which should be not_analyzed. Or, if you care about lowercase/uppercase, a keyword tokenizer together with a lowercase filter.
This should index Galapagos as is, not modifications. Then you can do your term search.
For example, a keyword analyzer together with lowercase filter:
"my_keyword_lowercase_analyzer": {
"type": "custom",
"filter": [
"lowercase"
],
"tokenizer": "keyword"
}
And the mapping:
"properties": {
"name":{
"type": "string",
"index_analyzer": "autocomplete",
"search_analyzer": "search_ngram",
"fields": {
"raw": {
"type": "string",
"analyzer": "my_keyword_lowercase_analyzer"
}
}
}
}
The query to be used is:
{
"query": {
"term": {
"name.raw": "galapagos islands"
}
}
}
So, instead of using the same field - name - you should be using name.raw (the subfield).