Elasticsearch illegal mapping of a location array

Question

I would like to perform geolocation queries with elasticsearch on my dataset which looks like this:

{"index": {"_index": "city","_type":"city_info", "_id":0} }
{"fields": { "city" : "AGAWAM", 
             "loc" : [ -72.622739, 42.070206 ], 
             "pop" : 15338, "state" : "MA"},
             "id" : "01001" ,
             "type" : "add"}
...

I am able to perform a lot of queries or aggregates without any problems but when it comes to a query with a geodistance filter or anything working on geographic coordinates, it doesn't work. Here is my mapping test:

PUT /city/_mappings
{
  "properties": {
    "fields.loc": {
      "type": "geo_point"
    }
  }
}

I get this error.

{
  "error" : {
    "root_cause" : [
      {
        "type" : "illegal_argument_exception",
        "reason" : "mapper [fields.loc] of different type, current_type [float], merged_type [geo_point]"
      }
    ],
    "type" : "illegal_argument_exception",
    "reason" : "mapper [fields.loc] of different type, current_type [float], merged_type [geo_point]"
  },
  "status" : 400
}

So it seems my "fields.loc" is a float whereas in the JSON it is an array with float values in. I tried to see what was the actual type of "loc" and it is indeed a float, which i don't understand why:

GET /city/_mapping

{
  "city" : {
    "mappings" : {
      "properties" : {
        "fields" : {
          "properties" : {
            "city" : {
              "type" : "text",
              "fields" : {
                "keyword" : {
                  "type" : "keyword",
                  "ignore_above" : 256
                }
              }
            },
            "loc" : {
              "type" : "float"
...

So I don't understant, or rather I don't how to modify "loc" from a float to a geo_point. As I am a beginner with elasticsearch, I am only using elasticsearch and kibana for cURL.

Jozef - Spatialized.io · Accepted Answer · 2020-03-16 14:55:28Z

1

Once you set the datatype implicitly to float (by bulk-syncing, as you probably did), it's gonna be difficult to convert the floats to geo_points. I'd recommend dropping the index, setting the correct mapping with the geo_point datatype and resyncing everything.

In case you'd like to go deeper, have a look at the _reindex api. Here's a quick tut. That's something you'll probably face when your system runs in a production environment where dropping is now an option.

FYI, using a custom _doc type is deprecated.

answered Mar 16, 2020 at 14:55

Jozef - Spatialized.io

17k4 gold badges29 silver badges79 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Axel JOLY Over a year ago

Thanks! I've done the first thing you advised me: deleting the index, setting a new index with a correct mapping having "loc" type being "geo_point") and bulk-syncing again. Now my geo_distance query is working

Jozef - Spatialized.io Over a year ago

Nice. Happy to help!

Collectives™ on Stack Overflow

Elasticsearch illegal mapping of a location array

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related