the elasticsearch index contains json as below, only relevant element is show
"geoLocation": {
"coordinates": [ [ -90.66487121582031, 42.49201965332031 ], [ -90.66487884521484, 42.49202346801758 ], [ -90.6648941040039, 42.492034912109375 ], [ -90.66490936279297, 42.49203872680664 ], [ -90.66492462158203, 42.492042541503906 ], [ -90.6649398803711, 42.49204635620117 ], [ -90.66495513916016, 42.49205017089844 ], [ -90.66497039794922, 42.4920539855957 ], [ -90.66498565673828, 42.492061614990234 ], [ -90.66500854492188, 42.492061614990234 ], [ -90.66502380371094, 42.49207305908203 ], [ -90.6650390625, 42.4920654296875 ] ],
"type": "linestring"
},
The template for generating the mapping is as below
PUT _template/template_1?include_type_name=true
{
"index_patterns": ["metromind-its-alerts-day2-*"],
"settings": {
"number_of_shards": 2
},
"mappings": {
"logs": {
"properties": {
"geoLocation": {
"type": "geo_shape"
}
}
}
}
}
the mapping generated is shown below
mapping showing the geoLocation type
When Kibana Maps are used it detects the geo_shape
Kibana Map to render Linestring
Note However no Linestring is rendered, please suggest the resolution
The line string is there, in Dubuque IL, it's just that it's extra small at the scale of the earth.
Just click on the following icon and Elastic Map will focus on it and you'll see it:
Related
I'm trying to insert a location that is using a circle processor to generate the circle points. Location [30, 10] in the document is generated with the circle points properly.
docs: Circle processor | Elasticsearch Guide [8.5] | Elastic
PUT _ingest/pipeline/polygonize_circles
{
"description": "translate circle to polygon",
"processors": [
{
"circle": {
"field": "circle",
"error_distance": 1,
"shape_type": "geo_shape"
}
}
]
}
PUT circles
{
"mappings": {
"properties": {
"circle": {
"type": "geo_shape"
}
}
}
}
PUT circles/_doc/2?pipeline=polygonize_circles
{
"circle": {
"type": "circle",
"radius": "40m",
"coordinates": [35.539917, -78.472000]
}
}
GET circles/_doc/2
But if I use another location. The generated coordinate looks like an oval with the wrong radius.
my location [35.54171753710938, -78.472]
created coordinates:
"circle": {
"coordinates": [
[
[
35.54171753710938,
-78.472
],
[
35.54112406581135,
-78.47173430472324
],
[
35.540630197847186,
-78.47167003953963
],
[
35.540375564960186,
-78.47165140797998
],
[
35.54021828908823,
-78.47164529506406
],
[
35.54010640465818,
-78.47164274491611
],
[
35.54001650261309,
-78.47164162397662
],
[
35.53993651647515,
-78.47164003979641
],
[
35.539858062238046,
-78.47164049555624
],
[
35.53977439153409,
-78.47164207973975
],
[
35.5396750942597,
-78.47164321572573
],
[
35.5395458932956,
-78.47164846905713
],
[
35.539348254773515,
-78.47165779735127
],
[
35.53899878746994,
-78.47169061817682
],
[
35.53833938849573,
-78.47182842440924
],
[
35.53833938849573,
-78.47217157559075
],
[
35.53899878746994,
-78.47230938182317
],
[
35.539348254773515,
-78.47234220264872
],
[
35.5395458932956,
-78.47235153094286
],
[
35.5396750942597,
-78.47235678427425
],
[
35.53977439153409,
-78.47235792026024
],
[
35.539858062238046,
-78.47235950444374
],
[
35.53993651647515,
-78.47235996020358
],
[
35.54001650261309,
-78.47235837602337
],
[
35.54010640465818,
-78.47235725508388
],
[
35.54021828908823,
-78.47235470493592
],
[
35.540375564960186,
-78.47234859202001
],
[
35.540630197847186,
-78.47232996046036
],
[
35.54112406581135,
-78.47226569527675
],
[
35.54171753710938,
-78.472
]
]
],
"type": "Polygon"
}
coordinates mapping on google maps
Is it an issue or It's working as expected? Because the coordinates are not a circle so it's impacting the search result.
You need to specify your coordinate array using longitude first and then latitude, I think you did the opposite and your circle is in the middle of Antartica.
If you do it like this:
PUT circles/_doc/2?pipeline=polygonize_circles
{
"circle": {
"type": "circle",
"radius": "40m",
"coordinates": [-78.472000, 35.539917]
}
}
Then your circle doesn't look oval anymore:
From the official doc:
In GeoJSON and WKT, and therefore Elasticsearch, the correct coordinate order is longitude, latitude (X, Y) within coordinate arrays. This differs from many Geospatial APIs (e.g., Google Maps) that generally use the colloquial latitude, longitude (Y, X).
How do I index a document with below data in elasticsearch(geo datatype)?
<west>5.8663152683722</west>
<north>55.0583836008072</north>
<east>15.0418156516163</east>
<south>47.2701236047002</south>
I tried geo_point and its working for lon and lat's, not sure how to save this data. any help is highly appreciated.
You'll have to use the geo_shape datatype and convert your XML (I assume) semi-points into a line string or polygon before sync.
I'm gonna go with a polygon here.
Let's visualise the conventional cardinal directions:
North (+90)
|
(-180) West ——+—— East (+180)
|
South (-90)
geo_shape expects GeoJSON-like inputs so you'll need five coordinate points, the first and last of which are identical (according to the GeoJSON spec).
Therefore, borrowing from TurfJS and going from bottom left counter-clockwise,
const lowLeft = [west, south];
const topLeft = [west, north];
const topRight = [east, north];
const lowRight = [east, south];
return
[
[
lowLeft,
lowRight,
topRight,
topLeft,
lowLeft
]
]
Finally, let's create our index and plug your numbers in
PUT /example
{
"mappings": {
"properties": {
"location": {
"type": "geo_shape"
}
}
}
}
POST /example/_doc
{
"location":{
"type":"polygon",
"coordinates":[
[
[
5.8663152683722,
47.2701236047002
],
[
15.0418156516163,
47.2701236047002
],
[
15.0418156516163,
55.0583836008072
],
[
5.8663152683722,
55.0583836008072
],
[
5.8663152683722,
47.2701236047002
]
]
]
}
}
Then verify that the center of your square if indeed inside of your indexed polygon:
GET example/_search
{
"query": {
"geo_shape": {
"location": {
"shape": {
"type": "point",
"coordinates": [
10.45406545999425,
51.1642536027537
]
},
"relation": "intersects"
}
}
}
}
I try to edit a geojson file to keep only objects that have the key "name".
The filter works but I can't find a way to keep the other objects and, specifically, the geometry and redirect the whole stuff to a new geojson file. Is there a way to display the whole object after filtering one of its children objects?
Here is an example of my data. The first object has the "name" property and the second hasn't:
{
"features": [
{
"type": "Feature",
"id": "way/24824633",
"properties": {
"#id": "way/24824633",
"highway": "tertiary",
"lit": "yes",
"maxspeed": "50",
"name": "Rue de Kleinbettingen",
"surface": "asphalt"
},
"geometry": {
"type": "LineString",
"coordinates": [
[
5.8997935,
49.6467825
],
[
5.8972561,
49.6467445
]
]
}
},
{
"type": "Feature",
"id": "way/474396855",
"properties": {
"#id": "way/474396855",
"highway": "path"
},
"geometry": {
"type": "LineString",
"coordinates": [
[
5.8020608,
49.6907648
],
[
5.8020695,
49.6906054
]
]
}
}
]
}
Here is what I tried, using jq
cat file.geojson | jq '.features[].properties | select(has("name"))'
The "geometry" is also a child of "features" but I can't find a way to make the selection directly from the "features" level. Is there some way to do that? Or a better path to the solution?
So, the required ouput is:
{
"type": "Feature",
"id": "way/24824633",
"properties": {
"#id": "way/24824633",
"highway": "tertiary",
"lit": "yes",
"maxspeed": "50",
"name": "Rue de Kleinbettingen",
"surface": "asphalt"
},
"geometry": {
"type": "LineString",
"coordinates": [
[
5.8997935,
49.6467825
],
[
5.8972561,
49.6467445
]
]}}
You can assign the filtered list back to .features:
jq '.features |= map(select(.properties|has("name")))'
I have following mapping in elastic search. I am able to PUT documents using Sense plugin but unable to do so using XContentBuilder to set the geo_shape field value. I am getting following error:
error:
[106]: index [streets], type [street], id [{dc872755-f307-4c5e-93f6-bba9c95791c7}], message [MapperParsingException[failed to parse [shape]]; nested: ElasticsearchParseException[shape must be an object consisting of type and coordinates];]
mapping:
PUT /streets
{
"mappings": {
"street": {
"properties": {
"id": {
"type": "string"
},
"shape": {
"type": "geo_shape",
"tree": "quadtree"
}
}
}
}
}
code:
val bulkRequest:BulkRequestBuilder = esClient.prepareBulk()
//inloop
xb = jsonBuilder().startObject()
xb.field("id", guid)
xb.field("shape", jsonString) // removing this line creates the index OK but without the geo_shape
xb.endObject()
bulkRequest.add(esClient.prepareIndex("streets", "street", guid).setSource(xb))
//end loop
val bulkResponse:BulkResponse = bulkRequest.execute().actionGet()
if(bulkResponse.hasFailures){
println(bulkResponse.buildFailureMessage())
}
jsonString:
{
"id": "{98b8fd8d-074c-4349-a83b-6e892bf2d0ef}",
"shape": {
"type": "LineString",
"coordinates": [
[-70.81866815832467, 43.12187109162505],
[-70.83054813653018, 43.15917412985851],
[-70.81320737213957, 43.23522269547419],
[-70.90108590067649, 43.28102004268419]
],
"crs": {
"type": "name",
"properties": {
"name": "EPSG:4326"
}
}
}
}
Appreciate any feedback?
Thanks
It might be a bit late for you, but this could help someone facing a similar issue even nowadays.
Following your index mapping for the document streets, we have these properties: id and shape.
In your error message, it's described that:
shape must be an object consisting of type and coordinates
So for your concrete case, the crs array is just not accepted (don't know exactly why you can't add extra parameters).
This is an example for how to add a document into the streets index using CURL:
curl -X POST "localhost:9200/streets/_doc?pretty" -H 'Content-Type: application/json' -d '
{
"id": 123,
"shape": {
"type": "Polygon",
"coordinates": [
[
[
32.85444259643555,
39.928694653732364
],
[
32.847232818603516,
39.9257985682691
],
[
32.837791442871094,
39.91947941109337
],
[
32.837276458740234,
39.91579296675271
],
[
32.85392761230469,
39.913423004886894
],
[
32.86937713623047,
39.91329133793421
],
[
32.88036346435547,
39.91539797880347
],
[
32.85444259643555,
39.928694653732364
]
]
]
}
}'
If you need to add a LineString, instead of a Polygon, just change the 'type' attribute from the 'shape'.
I hope this helps people having to add documents with shapes into an ElasticSearch database.
How is d3.nest() used with geojson files?
My geojson data is formatted as follows:
"features": [
{ "type": "Feature", "properties": { "neighborhood": "Allerton", "boroughCode": "2", "borough": "Bronx", "#id": "http:\/\/nyc.pediacities.com\/Resource\/Neighborhood\/Allerton" }, "geometry": { "type": "Polygon", "coordinates": [ [ [ -73.848597000000183, 40.871670000000115 ], [ -73.845822536836778, 40.870239076236174 ], [ -73.854559184633743, 40.859953835764252 ], [ -73.854665433068263, 40.859585694988056 ], [ -73.856388703358959, 40.857593635304482 ], [ -73.868881809153407, 40.857223150158326 ], [ -73.868317552728243, 40.857862062258313 ], [ -73.869553714672321, 40.857784095600181 ], [ -73.871024857620654, 40.857309948816905 ], [ -73.870480549987164, 40.865413584098484 ], [ -73.87055489856489, 40.869702798589863 ], [ -73.86721594442561, 40.869689663636713 ], [ -73.85745, 40.869533000000182 ], [ -73.855550000000108, 40.871813000000145 ], [ -73.853597967576576, 40.873288368674203 ], [ -73.848597000000183, 40.871670000000115 ] ] ] } }
But my nest command:
var nested_data = d3.nest()
.key(function(d, i) { console.log(d); return d.features.properties.neighborhood; })
.entries(map);
returns an empty array.
I want to nest my data to more easily filter it. Is this advised?
Assuming your geojson looks like the below
var map = {
type: "FeatureCollection",
"features": [
{
"type": "Feature",
"properties": {
"neighborhood": "Allerton",
"boroughCode": "2",
"borough": "Bronx",
"#id": "http:\/\/nyc.pediacities.com\/Resource\/Neighborhood\/Allerton"
},
"geometry": { /* various coordinates, etc */ }
]
}
So, what you want to do is:
d3.nest()
.key(function(d, i) {
return d.properties.neighborhood;
})
.entries(map.features);
You want to pass map.features since that's your array.