The last piece in the puzzle is to demonstrate how to use a dataset definition in an OpenSAFELY study.
An OpenSAFELY study consists of a set of actions.
At least one action must be an ehrQL action, to extract a dataset from an OpenSAFELY backend.

You can run a single action using opensafely exec.

In your Codespace, open a terminal by pressing Ctrl+J, and run:

opensafely exec ehrql:v1 generate-dataset dataset_definition.py --dummy-tables dummy_tables

You should see the terminal fill with a table of data in CSV format.
Scroll up to see the column headers, and notice the two columns from your dataset definition (prt_or_mal and ace_or_arb).

Question: what happens if you rename the dataset variable and run the opensafely exec command again?

??? tip "The anatomy of an OpenSAFELY command"
What do the parts of the OpenSAFELY command
opensafely exec ehrql:v1 generate-dataset dataset_definition.py
do?

* `opensafely exec` executes an OpenSAFELY action independently of other OpenSAFELY actions
* `ehrql` is the OpenSAFELY action to execute
* `v1` is the major version of the ehrQL action
* `generate-dataset` is the ehrQL command to generate a dataset from a dataset definition
* `dataset_definition.py` is the dataset definition
* `--dummy-tables dummy_tables` gives the path to the dummy data

Note: the main OpenSAFELY tutorial documents how you can describe the actions of your study in a file called project.yaml.

Next: Test your ehrQL knowledge with a quiz!

!!! abstract "Feedback"
Don't fancy the quiz? That's not a problem, but if you could fill in this very short feedback form{:target="_blank"} we'd really appreciate it.