Elasticsearch Mapping and Settings definitions - elasticsearch

Question

I tried setting a new index by configuring its mapping and settings.
Here is the code I used:
POST /test/text
{
"settings": {
"analysis": {
"filter": {
"greek_stop": {
"type": "stop",
"stopwords": "_greek_"
},
"greek_lowercase": {
"type": "lowercase",
"language": "greek"
},
"greek_stemmer": {
"type": "stemmer",
"language": "greek"
}
},
"analyzer": {
"greek": {
"tokenizer": "standard",
"filter": [
"greek_lowercase",
"greek_stop",
"greek_stemmer"
]
}
}
}
},
"mappings": {
"article": {
"properties": {
"title": {
"type": "string",
"fields": {
"greek": {
"type": "string",
"analyser": "greek"
}
}
},
"content": {
"type": "string",
"fields": {
"greek": {
"type": "string",
"analyser": "greek"
}
}
},
"indexed_date": {
"type": "date",
"format": "strict_date_optional_time||epoch_millis"
}
}
}
}
}
I used POST as I didn't care much about the _id and that way ES gives a random value to this variable instead of using PUT where I have to specify its value. The reason why I created two fields for title and content is because I wish to have the raw version of text and the 'stop words removed-stemmed' version in order to be able to weight higher if the term is found exactly as the user entered it (instead of storing only the stemmed version of a word)
After populating the index with data, e.g.:
PUT /test/text
{
"title": " ",
"content": " ",
"date": " "
}
I tried doing a search query like this:
GET /test/text/_search
{
"query":{
"multi_match":{
"query":"όμορφος",
"type":"most_fields",
"fields":["content","content.greek","title","title.greek"]}}}
And then by changing the query to "όμορφη", these words have the same stemmed version which is "όμορφ", and as a result due to the ES's greek language analyzer I should get the same entry, which I don't.
Any idea why? Should I be doing something more while indexing my documents? After reading the documentation I was under the impression that after defining the mapping when indexing a piece of text it would automatically get indexed both ways and the query would each time be analysed using the appropriate analyzer automatically.
If I am under the right impression why my query doesn't return the same results? Any ideas?
Thank you in advance.

Answer 1

To create the index in the first place you need to call PUT test not POST /test/text. The latter will simply create a new document of type text in a new index called test, but with the default settings and mappings.
So first:
PUT /test
{
"settings": {
...
},
"mappings": {
...
}
}
Then you can create new documents with (note that your mapping type was called article, not text)
POST /test/article
{
"title": " ",
"content": " ",
"date": " "
}
Then only your search query will work