Supposing that the MongoDB container is running and we possess a CSV file that contains records that constitute spatial points (longitude, latitude values). We proceed to bulk insertion by the following steps:
- Copy the CSV file to the MongoDB container:
$docker cp mycsv.csv mongodb:/mycsv.csv
- Connect to the MongoDB container:
$sudo docker exec -it mongodb bash
- Run the importing command:
$mongoimport --db=test --collection=points --type=csv --file=./mycsv.csv --host=localhost:27017 --columnsHaveTypes --fields="longitude.double(), latitude.double()"
The CSV records will be inserted as documents with two double type fields, named longitude and latitude. The records will be inserted in points collection of test database.
By default, the batch size of mongoimport tool for bulk insertion is set to 100,000
- To represent the longitude and latitude as points, we transform the fields in GeoJSON objects (this command should run on the mongo shell – run mongo ), named location:
$db.points.updateMany({}, [{"$set":{ "location" : {type:"Point", coordinates:["$longitude", "$latitude"] } } } ] )
- Delete the longitude and latitude fields, as we have formed the GeoJSON objects:
$db.points.updateMany({}, {"$unset":{longitude:1, latitude:1 }})