Elasticsearch Bulk Insert JSON Data

Onexlab

Feb 11·6 min read

In this article, We will show you how to insert bulk JSON data to Elasticsearch.

We hope you are familiar with Elasticsearch if not, don’t worry you will learn from this article.

Video

Initial Setup

The first step is to setup Elasticsearch on your system. We are using Docker to setup. Make sure you have installed docker on your system or can download it from the Docker webpage.

Let’s start

Create a new file docker-compose.yml under any directory, In our case, we are using the following project structure.

Project (Directory)
├── docker-compose.yml (File)

docker-compose.yml

In the above file, we are creating 2 containers one for Elasticsearch and the second for Kibana

To start containers, Run the following command in the terminal

docker-compose up

When you run the above command you will see the following output in the terminal.

Verify Running Elasticsearch

To verify, Running container of Elasticsearch. Open the browser and type the following URL

http://localhost:9200

You will see the JSON response in the browser as shown below in the screenshot.

Verify Running Kibana

To verify, Running container of Kibana. Open the browser and type the following URL

http://localhost:5601

You will see the Welcome screen in the browser as shown below in the screenshot.

We are done with Elasticsearch and Kibana setup using Docker Compose

Create Elasticsearch Index

If you are familiar with relational databases such as MySql and Postgres. If you want to create a new table you need to create the database first then you can add a table to that database. Similar to Elasticsearch, we create a document instead of a table because Elasticsearch is a NoSQL database and you can’t create the document without an index have a look at the following table comparison between MySql and Elasticserach

To create a new index use the following CURL

PUT (Method Type)
http://localhost:9200/cart?&prettyCURL:
curl --request PUT \
  --url 'http://localhost:9200/cart?pretty=' \
  --header 'Connection: keep-alive'

Using the above CURL, we have created the index name cart as shown below in the screenshot, We used Insomnia for testing you can download and try

Ingest Single JSON Object

We have created the cart index before, Now we will ingest a Single JSON document in the index cart

{
 "id": 1,
 "title":"iPhone 11",
 "category": "Mobile Phone",
 "price": 123300,
 "formatted_price": "$1233.00 USD",
 "description": "iPhone 11, 64 GB "
}

To insert a Single JSON object you can use the following CURL with the above JSON body

PUT (Method Type)
http://localhost:9200/cart/_doc/?prettyCURL:
curl --request POST \
  --url 'http://localhost:9200/cart/_doc/?pretty=' \
  --header 'Content-Type: application/json' \
  --data '{
 "id": 1,
 "title":"iPhone 11",
 "category": "Mobile Phone",
 "price": 123300,
 "formatted_price": "$1233.00 USD",
 "description": "iPhone 11, 64 GB "
}'

Using the above CURL, we have stored the Single JSON document as shown below in the screenshot

Elasticsearch Mapping

As we have data types in the Relation Databases, We do have data types in the Elasticsearch as well called Mapping. In the above screenshot, we have inserted a Single JSON object to the Elasticsearch. If you noticed we haven’t created any Mapping before. It was created automatically. If you want to check the Mapping you can use the following CURL

GET (Method Type)
http://localhost:9200/cart/_mapping?prettyCURL:
curl --request GET \
  --url 'http://localhost:9200/cart/_mapping?pretty=' \
  --header 'Content-Type: application/json'

Using the above CURL, You can see the document Mapping as shown below in the screenshot

Elasticsearch Get All Document

We have inserted a Single JSON object before. To check an inserted document in the index cart. Use the following CURL with JSON Body

GET (Method Type)
http://localhost:9200/cart/_search?prettyJSON Body
{
 "query": {
  "match_all": {
  }
 }
}CURL:
curl --request GET \
  --url 'http://localhost:9200/cart/_search?pretty=' \
  --header 'Content-Type: application/json' \
  --data '{
 "query": {
  "match_all": {
  }
 }
}'

Using the above CURL, we can see all the stored documents in the index cart as shown below in the screenshot

Elasticsearch Ingest Bulk/Multiple Documents

Insert Single JSON document is very simple in the Elasticsearch as compare to the Multiple/Bulk data ingestion. You need to convert the following JSON to NDJSON format

JSON
{
 "id": 1,
 "title":"iPhone 11",
 "category": "Mobile Phone",
 "price": 123300,
 "formatted_price": "$1233.00 USD",
 "description": "iPhone 11, 64 GB "
}NDJSON
{"index":{"_index":"cart","_id":1,"_type":"_doc"}}
{"id":1,"title":"iPhone 11","category":"Mobile Phone","price":123300,"formatted_price":"$1233.00 USD","description":"iPhone 11, 64 GB "}

There is a very nice library called json-to-es-bulk you can use to generate the NDJSON format bulk data for Elasticsearch

We have experimented on Repl.it. You can also use that to convert to NDSON

We have the following test.json file and want to convert it into NDJSON format to insert to Elasticsearch using the following command under the Shell Tab on Repl.it

node index.js -f test.json --index cart --type _doc

Try to use the Shell tab as shown in the following screenshot

It will generate a new file request-data.txt contains NDJSON formatted JSON data as shown below

We can use the above data to insert bulk JSON documents to the Elasticsearch

Let’s use the following CURL to ingest bulk JSON data to the Elasticsearch

POST (Method Type)
http://localhost:9200/cart/_bulk/?prettyCURL:
curl --request POST \
  --url 'http://localhost:9200/cart/_bulk/?pretty=' \
  --header 'Content-Type: application/x-ndjson' \
  --data '{"index":{"_index":"cart","_id":1,"_type":"_doc"}}
{"id":1,"title":"iPhone 11","category":"Mobile Phone","price":123300,"formatted_price":"$1233.00 USD","description":"iPhone 11, 64 GB "}
{"index":{"_index":"cart","_id":2,"_type":"_doc"}}
{"id":2,"title":"iPhone 11 Pro","category":"Mobile Phone","price":143300,"formatted_price":"$1433.00 USD","description":"iPhone 11 Pro, 128 GB "}
{"index":{"_index":"cart","_id":3,"_type":"_doc"}}
{"id":3,"title":"iPhone X","category":"Mobile Phone","price":103300,"formatted_price":"$1033.00 USD","description":"iPhone X, 128 GB "}
'

When you use the above CURL it will ingest bulk JSON data as shown below in the screenshot

To verify, the ingested JSON data you can you the following CURL

GET (Method Type)
http://localhost:9200/cart/_search?prettyCURL:
curl --request GET \
  --url 'http://localhost:9200/cart/_search?pretty=' \
  --header 'Content-Type: application/json' \
  --data '{
 "query": {
  "match_all": {
  }
 }
}'

Using the above CURL, you can retrieve all the ingested documents as shown below on the screenshot

We are done with the bulk JSON data Ingestion with Elasticsearch

Thank you! 🚀🚀🚀🚀🚀🚀