How do I test the sensor cursor has been updated?

To test that a sensor's cursor has been updated, you can follow these steps: 1. Via the Dagster UI: After manually triggering a test evaluation of a sensor in the UI, you can observe the new computed cursor value from the evaluation. If the cursor has been updated, the new value will be displayed with the option to persist it. 2. Via Python: When unit testing sensors, you can directly invoke the sensor's Python function and check the cursor value. Here's an example provided in the documentation:

python     from dagster import build_sensor_context          def test_my_directory_sensor_cursor():         context = build_sensor_context(cursor="0")         for run_request in my_directory_sensor_cursor(context):             # Your assertions or checks for the cursor update             # For example, you might check context.cursor here

In this example, you would replace

my_directory_sensor_cursor

with your actual sensor function. After invoking the sensor function, you can check the

context.cursor

to see if it has been updated as expected. Remember that when you test sensors via Python, you can also provide the necessary resources if your sensor uses them. This allows you to fully simulate the sensor's execution environment and verify that the cursor is being updated correctly. Keep in mind that any side effects contained within the sensor's function may be executed during these tests, so you should ensure that the test environment is isolated from production systems to avoid unintended consequences.

@Scout It doesn't seem I can access

context.sensor

after the execution is finished tho?

You are correct that after the execution of a sensor function, you cannot directly access the

context.sensor

to check the updated cursor value. The

context

object provided to the sensor function is used during the execution of the sensor to access information and perform actions, but it does not persist state after the function completes. To test that the cursor has been updated, you would typically assert the expected behavior within the sensor function itself or check the persistent state where the cursor is stored after the sensor execution. In Dagster, the cursor is stored in the DagsterInstance, which manages the state and execution of pipelines and sensors. If you want to check that the cursor has been updated as part of a unit test, you can do the following: 1. Use the

instance

that you provided to the

build_sensor_context

function to check the updated cursor value after the sensor function has been invoked. The

instance

represents the DagsterInstance, and it should have methods to retrieve the state of sensors, including the cursor. 2. If you are using the

DagsterInstance.ephemeral()

method to create an in-memory instance for testing, you can retrieve the sensor state from this instance after the sensor function has been invoked. Here's a conceptual example based on the documentation:

python from dagster import build_sensor_context, DagsterInstance  def test_my_sensor():     instance = DagsterInstance.ephemeral()     context = build_sensor_context(instance=instance)     my_sensor(context)     # After invoking the sensor, check the updated cursor value     # This will depend on the methods available on the DagsterInstance     # to retrieve the sensor state, which may include the cursor value.

Please note that the exact method to retrieve the sensor state and cursor from the

DagsterInstance

will depend on the Dagster version and API. The documentation does not provide a specific method call for this, so you would need to refer to the appropriate version of the Dagster API documentation or source code to find the correct method to use. If you are looking for a specific method to retrieve the cursor from the instance and the documentation does not provide this information, it may be helpful to look at the source code of Dagster or ask for guidance in the Dagster community or support channels.