elasticsearch: special behaviour of _id field? - mongodb

Question

I have some twitter data I want to work with. I want to be able to search for a name. When trying to generate ngrams of the 'name' and '_id' I run into some troubles.
first, I created the analyzers:
curl -XPUT 'localhost:9200/twitter_users' -d '
{
"settings": {
"analysis": {
"analyzer": {
"str_search_analyzer": {
"tokenizer": "keyword",
"filter": [
"lowercase"
]
},
"str_index_analyzer": {
"tokenizer": "keyword",
"filter": [
"lowercase",
"ngram"
]
}
},
"filter": {
"ngram": {
"type": "ngram",
"min_gram": 3,
"max_gram": 20
}
}
}
}
}'
then I defined my mappings:
curl -XPUT 'http://localhost:9200/twitter_users/users/_mapping' -d '
{
"users": {
"type" : "object",
"properties": {
"_id": {
"type": "string",
"copy_to": "id"
},
"id": {
"type": "string",
"search_analyzer": "str_search_analyzer",
"index_analyzer": "str_index_analyzer",
"index": "analyzed"
},
"name": {
"type": "multi_field",
"fields": {
"name": {
"type": "string",
"index": "not_analyzed"
},
"ngrams": {
"type": "string",
"search_analyzer": "str_search_analyzer",
"index_analyzer": "str_index_analyzer",
"index": "analyzed"
}
}
}
}
}
}'
and inserted some test data:
curl -XPUT "localhost:9200/twitter_users/users/johndoe" -d '{
"_id" : "johndoe",
"name" : "John Doe"
}'
curl -XPUT "localhost:9200/twitter_users/users/janedoe" -d '{
"_id" : "janedoe",
"name" : "Jane Doe"
}'
querying by name gets me the expected results:
curl -XPOST "http://localhost:9200/twitter_users/users/_search" -d '{
"query": {
"match": {
"name.ngrams": "doe"
}
}
}'
but querying on the id gives me no results:
curl -XPOST "http://localhost:9200/twitter_users/users/_search" -d '{
"query": {
"match": {
"id": "doe"
}
}
}'
I also tested to make _id a multi field like I did with name. But that didn't work either.
is _id behaving differently than other fields? Or am I doing something wrong here?
edit: using elasticsearch v1.1.2 and pulling the data from mongodb with a river plugin.
Thanks for your Help
Mirko

Answer 1

Looks like the 'copy_to' is the issue, but why not insert the 'id' values into the 'id' fields directly?
curl -XPUT "localhost:9200/twitter_users/users/johndoe" -d '{
"id" : "johndoe",
"name" : "John Doe"
}'
curl -XPUT "localhost:9200/twitter_users/users/janedoe" -d '{
"id" : "janedoe",
"name" : "Jane Doe"
}'