kt-search

Multi platform kotlin client for Elasticsearch & Opensearch with easily extendable Kotlin DSLs for queries, mappings, bulk, and more.

---

Project maintained by jillesvangurp Hosted on GitHub Pages — Theme by mattgraham

Document Manipulation

KT Search Manual	Previous: Deleting by query	Next: Index Repository
Github	© Jilles van Gurp

---

Mostly, you will use bulk indexing to manipulate documents in Elasticsearch. However, sometimes it is useful to be able to manipulate individual documents with the Create, Read, Update, and Delete (CRUD) APIs.

// create
val resp = client.indexDocument(
  target = "myindex",
  document = TestDoc("1", "A Document"),
  // optional id, you can let elasticsearch assign one
  id = "1",
  // this is the default
  // fails if the id already exists
  opType = OperationType.Create
)

// read
val doc = client.getDocument("myindex", resp.id)
  // source is a JsonDoc, which you can deserialize
  // with an extension function
  .source.parse<TestDoc>()

// update
client.indexDocument(
  target = "myindex",
  document = TestDoc("1", "A Document"),
  id = "1",
  // will overwrite if the id already existed
  opType = OperationType.Index
)

// delete
client.deleteDocument("myindex", resp.id)

Elasticsearch also has a dedicated update API that you can use with either a partial document or a script.

client.indexDocument(
  target = "myindex",
  document = TestDoc("42", "x"),
  id = "42"
)
var resp = client.updateDocument(
  target = "myindex",
  id = "42",
  docJson = """{"name":"changed"}""",
  source = "true"
)
println(resp.get?.source)

resp = client.updateDocument(
  target = "myindex",
  id = "42",
  script = Script.create {
    source = """ctx._source.name = params.p1 """
    params = mapOf(
      "p1" to "again"
    )
  },
  source = "true"
)
println(resp.get?.source)

Captured Output:

{"id":"42","name":"changed","tags":[]}
{"id":"42","name":"again","tags":[]}

The index API has a lot more parameters that are supported here as well via nullable parameters. You can also use a variant of the index API that accepts a json String instead of the TestDoc.

Note, for inserting large amounts of documents you should of course use the bulk API. You can learn more about that here: Efficiently Ingest Content Using Bulk Indexing.

---

KT Search Manual	Previous: Deleting by query	Next: Index Repository
Github	© Jilles van Gurp