ElasticSearch: Updating nested document in array

Question

Given the document

curl -XPUT 'localhost:9200/test/me/here' -d '{
  "top" : [
    { "searchkey" : "change"},
    { "searchkey" : "keep"}
  ]
}'

I need an update query that will add new field to sub-document with searchkey equal to change and will keep any other sub-document intact. The expected result is then:

{
  "top" : [
    { "searchkey" : "change", "newfield" : "newvalue"},
    { "searchkey" : "keep"}
  ]
}

Running query that selects inner document by index works, but I do not know the inner order in advance and it is quite fragile anyways:

curl -XPOST 'localhost:9200/test/me/here/_update' -d '{
    "script" : "ctx._source.top[0].newfield = v",
    "params" : {
      "v" : "newvalue"
    }
}'

Is there a way to tell ES to add the new field to the inner document that matches some condition? Something like:

curl -XPOST 'localhost:9200/test/me/here/_update' -d '{
    "script" : "ctx._source.top[ctx._source.top.searchkey == s].newfield = v",
    "params" : {
      "s" : "change",
      "v" : "newvalue"
    }
}'

Or, will I do better and save some headache if I eliminate the array and transform the document to:

{
"change" : {},
"keep" : {}
}

Have you looked at stackoverflow.com/questions/18360225/… , — progrrammer
– progrrammer, Commented Aug 18, 2014 at 11:32
Yes. That is not my case, I want to add new info to existing nested document which is itself in array. — Martin
– Martin, Commented Aug 18, 2014 at 11:34
Yaa, I got it, you have to reindex that document with updated value. — progrrammer
– progrrammer, Commented Aug 18, 2014 at 11:39

Robert Fišer · Accepted Answer · 2015-12-07 08:11:23Z

3

You can use the update with script. See example:

PUT test/data/3/
{
   "source": [
     {
       "name": "A",
       "count": 1
     },
     {
       "name": "B",
       "count": 2
     },
     {
       "name": "c",
       "count": 3
     }
   ]
}

GET test/data/3

POST test/data/3/_update
{
 "script": " for (int i = 0; i < source.size(); i++) {boolean f = false;for (int j = 0; j < ctx._source.source.size(); j++) {if (ctx._source.source[j].name == source[i].name) {ctx._source.source[j].count = source[i].count;f=true;break;}}\nif(!f){ctx._source.source.add(source[i]);}}",
 "params": {
   "source": [
     {
       "name": "A",
       "count": 10
     },
     {
       "name": "B",
       "count": 30
     },
     {
       "name": "D",
       "count": 50
     }
   ]
 }
}

GET test/data/3

answered Dec 7, 2015 at 8:11

Robert Fišer

2191 silver badge10 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Shyamala Rao · Accepted Answer · 2020-04-22 04:51:59Z

2

We can use update_by_query API to get the documents that matches the condition and then scroll through the array of json to update the object within the array.

POST test/data/_update_by_query
{
    "script": {
        "lang":"painless",
        "source":"""
            for(int i=0;i<ctx._source.top.length;i++){
                if(ctx._source.top[i].searchkey == params.match_value){
                    ctx._source.top[i].new_field = params.new_value;
                }
            }
        """,
        "params" : {
             "match_value" : "change",
             "new_value" : "new_value"
         }
    },
    "query" : {
        "match" : {
            "top.searchkey": "change"
        }
    }

}

answered Apr 22, 2020 at 4:51

Shyamala Rao

211 bronze badge

1 Comment

Yair Over a year ago

Thanks Shyamala Rao If you get an for error invalid content-type due to multi-line script you should use single quotes and single line, e.g.

"source":"for(int i=0;i<ctx._source.top.length;i++){ if(ctx._source.top[i].searchkey == params.match_value){ ctx._source.top[i].new_field = params.new_value; } }"

Collectives™ on Stack Overflow

ElasticSearch: Updating nested document in array

2 Answers 2

Comments

1 Comment

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Linked

Related