Providing field mapping, analysis, etc on the fly

Elastic Stack Elasticsearch
1

Hi all,

Is it possible to provide the fields mapping, analysis, etc in the payload itself ? What i mean is, assuming i have the JSON payload with fields in it

{
    "account_number": 0,
    "balance": 16623,
    "firstname": "Bradshaw",
    "lastname": "Mckenzie",
    "age": 29,
    "gender": "F",
    "address": "244 Columbus Place",
    "employer": "Euron",
    "email": "bradshawmckenzie@euron.com",
    "city": "Hobucken",
    "state": "CO"
}

Is it possible to provide metadata about fields like

"firstname" : {
       "type" : "string",
       "value" : "Bradshaw",
       "analysis": {
       "analyzer": {
        "std_folded": { 
          "type": "custom",
          "tokenizer": "standard",
          "filter": [
            "lowercase",
            "asciifolding"
           ]
         }
       }
     }
  }

with the payload rather than providing mapping on the index level ? Reason i am asking is, our schema wouldn't be well defined as all the documents/payload will have randoms field names and counts. Field Name in one document could be int and long in other document.

Any help or pointer is much appreciated.

Regards,
Lukes

2

Can anyone please provide some input ?

Regards,
Lukes

3

Each field in an index must have a single mapping, so it is not possible to set this per document.

4

Thanks @Christian_Dahlqvist. Then i'll put the naming conventions in my fields itself.

Regards,
Lukes

5

Is it possible to have wildcard type fields mapping ? What i mean is, like in SOLR we can define dynamicFields mapping saying map *_txt_en to text_en and then we define text_en as text field with english analyzer and so on. I tried

{
  "mappings": {
    "type1": {
      "properties": {
        "*_i_en": {
          "type": "integer"
        }
      }
    }
  }
}

and tried indexing below field
{
"age_i_en":25
}

I checked metadata and it was indexed as long, but not as integer. What i am trying to achieve is define wildcard/regular expression type mappings, so that i can define customize pipelines and so on. And during runtime fields names will get resolved automatically and this way i don't have to define/tie fields names. Also during query time is it possible to specify "age_i_*" as query field rather than complete name, so that if i have fields with multiple languages then i don't have to iterate over each of them ?

Regards,
Lukes

6

Hi experts, can anyone help please ?

Regards,
Lukes

7

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.