CODEWITHSUNDEEP
elasticsearchkibana
Maunil Vyas
Maunil Vyas

Reputation: 85

Removing an existing field from an index in Elastic Search

I am a newbie in Elastic search. I have an unwanted field in my index say index name "test_index". This contains almost 155154 documents. I want to remove an unwanted field "B" from my index. This is how my index pattern look like in json format

{A : {B : {C: } }

I believe removing B will automatically remove C from my index as well. In order to do that, I used the following query but it didn't seem to work. 

POST test_index/_update_by_query?conflicts=proceed {
      "script" : "ctx._source.A.remove('B')", 
       "query" : {
        "exists": { "field": "A.B" }
      }
}

Please let me know where I am doing the mistake.

Thank you

Upvotes: 0

Views: 2003

Answers (1)

Polynomial Proton
Polynomial Proton

Reputation: 5135

Your syntax is correct. You get timeout since its running the process in the background and times out before completing the task.

You can run the query asynchronously by specifying wait_for_completion=false

POST test_index/_update_by_query?conflicts=proceed&wait_for_completion=false
{
      "script" : "ctx._source.A.remove('B')", 
       "query" : {
        "exists": { "field": "A.B" }
      }
}

Above will give a response with taskId

{
  "task" : "{taskId:node}"
}

Now you can use task api to get the status of the task using the value from above

GET _tasks/{taskId:node}

Alternatively, if you dont specify wait_for_completion=false and get a time out, you can still get all tasks by actions like below. But, i'd recommend do the first one.

GET _tasks?actions=*byquery&detailed.

From comments: Now say if I have 100 index having a similar name pattern for example an index having a name "test_date - MM/DD/YYYY" here the prefix is same "test"

In order to handle multiple indices, you can use the wild card syntax and replace index name with prefix and *

For example, below query will run on all indices that start with test:

POST test*/_update_by_query?conflicts=proceed&wait_for_completion=false
{
      "script" : "ctx._source.A.remove('B')", 
       "query" : {
        "exists": { "field": "A.B" }
      }
}

Upvotes: 3

Related Questions