Skip to main content
Version: 0.12.1

Getting Started With GraphQL

Reading an Entity: Queries

DataHub provides the following graphql queries for retrieving entities in your Metadata Graph.

Query

The following graphql query retrieves the urn and name of the properties of a specific dataset

{
dataset(urn: "urn:li:dataset:(urn:li:dataPlatform:kafka,SampleKafkaDataset,PROD)") {
urn
properties {
name
}
}
}

In addition to the URN and properties, you can also fetch other types of metadata for an asset, such as owners, tags, domains, and terms of an entity. For more information on, please refer to the following links."

To perform full-text search against an Entity of a particular type, use the search(input: SearchInput!) graphql Query. The following graphql query searches for datasets that match a specific query term.

{
search(input: { type: DATASET, query: "my sql dataset", start: 0, count: 10 }) {
start
count
total
searchResults {
entity {
urn
type
...on Dataset {
name
}
}
}
}
}

The search field is used to indicate that we want to perform a search. The input argument specifies the search criteria, including the type of entity being searched, the search query term, the start index of the search results, and the count of results to return.

The query term is used to specify the search term. The search term can be a simple string, or it can be a more complex query using patterns.

  • * : Search for all entities.
  • *[string] : Search for all entities that contain aspects starting with the specified [string].
  • [string]* : Search for all entities that contain aspects ending with the specified [string].
  • *[string]* : Search for all entities that match aspects named [string].
  • [string] : Search for all entities that contain the specified [string].
note

Note that by default Elasticsearch only allows pagination through 10,000 entities via the search API. If you need to paginate through more, you can change the default value for the index.max_result_window setting in Elasticsearch, or using the scroll API to read from the index directly.

Modifying an Entity: Mutations

note

Mutations which change Entity metadata are subject to DataHub Access Policies. This means that DataHub's server will check whether the requesting actor is authorized to perform the action.

To update an existing Metadata Entity, simply use the update<entityName>(urn: String!, input: EntityUpdateInput!) GraphQL Query. For example, to update a Dashboard entity, you can issue the following GraphQL mutation:

mutation updateDashboard {
updateDashboard(
urn: "urn:li:dashboard:(looker,baz)",
input: {
editableProperties: {
description: "My new desription"
}
}
) {
urn
}
}

For more information, please refer to following links.

Please refer to Datahub API Comparison to navigate to the use-case oriented guide.

Handling Errors

In GraphQL, requests that have errors do not always result in a non-200 HTTP response body. Instead, errors will be present in the response body inside a top-level errors field.

This enables situations in which the client is able to deal gracefully will partial data returned by the application server. To verify that no error has returned after making a GraphQL request, make sure you check both the data and errors fields that are returned.

To catch a GraphQL error, simply check the errors field side the GraphQL response. It will contain a message, a path, and a set of extensions which contain a standard error code.

{
"errors": [
{
"message": "Failed to change ownership for resource urn:li:dataFlow:(airflow,dag_abc,PROD). Expected a corp user urn.",
"locations": [
{
"line": 1,
"column": 22
}
],
"path": ["addOwners"],
"extensions": {
"code": 400,
"type": "BAD_REQUEST",
"classification": "DataFetchingException"
}
}
]
}

With the following error codes officially supported:

CodeTypeDescription
400BAD_REQUESTThe query or mutation was malformed.
403UNAUTHORIZEDThe current actor is not authorized to perform the requested action.
404NOT_FOUNDThe resource is not found.
500SERVER_ERRORAn internal error has occurred. Check your server logs or contact your DataHub administrator.

Visit our Slack channel to ask questions, tell us what we can do better, & make requests for what you'd like to see in the future. Or just stop by to say 'Hi'.