Extract some value without key from a json object - json

Question

I am using java to extract a list of node ids from an elastic search running tasks response.
The response looks like this
{
"nodes": {
"ZKUuxQZpRCCcJ0njBM1P0A": {
"name": "ZKUuxQZ",
"transport_address": "127.0.0.1:9300",
"host": "127.0.0.1",
"ip": "127.0.0.1:9300",
"roles": [
"master",
"data",
"ingest"
],
"tasks": {
"ZKUuxQZpRCCcJ0njBM1P0A:118": {
"node": "ZKUuxQZpRCCcJ0njBM1P0A",
"id": 118,
"type": "transport",
"action": "indices:data/write/delete/byquery",
"start_time_in_millis": 1527808643421,
"running_time_in_nanos": 154234724059,
"cancellable": true,
"headers": {}
}
}
}
}
}
In this example, I want to exact ZKUuxQZpRCCcJ0njBM1P0A:118. Can someone give me an example how to extract this information?
Option 1, java json parser. I wanted to write a response class and parse the json string, but ZKUuxQZpRCCcJ0njBM1P0A:118 is not in the key:value format. I am not sure how to extract it.
Option 2, using regex. The node id part may also have - or _. Can someone provide a neat java regex solution? So far my regex is [a-zA-Z0-9\-\_]+\:[0-9]+, I am not sure if this is safe enough.
Thanks.

Answer 1

One way you can do this is by using org.json library and create JSONObject and using keys() you can get all the keys and use this to achieve your task
Add the following dependency in your pom.xml
<dependency>
<groupId>org.json</groupId>
<artifactId>json</artifactId>
<version>20180130</version>
</dependency>
Then you can do something like below:
String jsonString = "{\n" +
" \"nodes\": {\n" +
" \"ZKUuxQZpRCCcJ0njBM1P0A\": {\n" +
" \"name\": \"ZKUuxQZ\",\n" +
" \"transport_address\": \"127.0.0.1:9300\",\n" +
" \"host\": \"127.0.0.1\",\n" +
" \"ip\": \"127.0.0.1:9300\",\n" +
" \"roles\": [\n" +
" \"master\",\n" +
" \"data\",\n" +
" \"ingest\"\n" +
" ],\n" +
" \"tasks\": {\n" +
" \"ZKUuxQZpRCCcJ0njBM1P0A:118\": {\n" +
" \"node\": \"ZKUuxQZpRCCcJ0njBM1P0A\",\n" +
" \"id\": 118,\n" +
" \"type\": \"transport\",\n" +
" \"action\": \"indices:data/write/delete/byquery\",\n" +
" \"start_time_in_millis\": 1527808643421,\n" +
" \"running_time_in_nanos\": 154234724059,\n" +
" \"cancellable\": true,\n" +
" \"headers\": {}\n" +
" }\n" +
" }\n" +
" }\n" +
" }\n" +
"}";
JSONObject jsonObject = new JSONObject(jsonString);
Set<String> topLevelKeys = jsonObject.keySet();
for (String key : topLevelKeys) {
Object value = jsonObject.get(key);
if (value instanceof JSONObject) {
JSONObject valueObject = jsonObject.getJSONObject(key);
System.out.println(valueObject.toString());
}
}
Extend this as per your requirement.

Answer 2

So regex is too hacky and I figured it out using gson. I wish ElasticSearch can offer us some standard library to extract all kinds of responses. Here is my solution using gson.
import com.google.gson.*;
import org.json.JSONObject;
public class test {
public static void main(String[] args) throws Exception {
String jsonString = "json_format elasticsearch reponse for listing running tasks"
JsonParser parser = new JsonParser();
JsonObject jsonObject = parser.parse(content).getAsJsonObject();
jsonObject.getAsJsonObject("nodes").entrySet().forEach(
s -> {
s.getValue().getAsJsonObject().getAsJsonObject("tasks").entrySet().forEach( s2 -> {
System.out.println(s2.getKey());
});
});
}
It prints all the running task ids like the following. It's technically nodeId:taskId, but ES has a very vague description in their task API docs(It says you can use TASK_ID to check a status of a task, well that TASK_ID is nodeId:taskId).
Mw-3i39gTHGxu5c8z9viQQ:503209021
DZ29LMsWR0aW9guWZTYe2Q:482931604
6CAbDZSWR8SfwZgnRT0qNg:494351185

Answer 3

Use the following regex:
[a-zA-Z0-9-_]*[0-9]+[a-zA-Z]+:[0-9]*
You can find the test result here:
https://regexr.com/3qdug