Pushing and pulling XCom values in Apache Airflow - Practice Problems & Coding Challenges

Challenge - 5 Problems

💻 Command Output

intermediate

What is the output of pulling an XCom value?

Given a task pushes an XCom value with key 'result' and value 42, what will be the output of pulling this XCom in another task?

Apache Airflow

value = ti.xcom_pull(task_ids='push_task', key='result')
print(value)

A None

B '' (empty string)

C KeyError

D 42

💡 Hint

Remember that xcom_pull returns the value pushed with the matching key and task_id.

The ti.xcom_pull method retrieves the value pushed by the specified task and key. Since the value 42 was pushed with key 'result', pulling it returns 42.

🧠 Conceptual

intermediate

Which method is used to push an XCom value in Airflow?

You want to share data from one task to another using XComs. Which method should you use inside a task to push a value?

A ti.xcom_push(key='my_key', value='data')

B push_xcom(ti, key='my_key', value='data')

C xcom_push(key='my_key', value='data')

D ti.push_xcom(key='my_key', value='data')

💡 Hint

The method is called on the task instance object (ti).

The correct method to push an XCom value is ti.xcom_push(). It takes a key and a value to store.

❓ Troubleshoot

advanced

Why does xcom_pull return None unexpectedly?

You have two tasks: 'push_task' pushes an XCom with key 'data'. 'pull_task' tries to pull it but gets None. What is a likely cause?

A 'pull_task' runs before 'push_task' completes

B The key used in xcom_pull is misspelled

C All of the above

D 'push_task' did not call ti.xcom_push

💡 Hint

Consider timing, spelling, and whether the push actually happened.

All these reasons can cause ti.xcom_pull to return None: pulling before push, wrong key, or no push call.

🔀 Workflow

advanced

Order of operations for pushing and pulling XComs

Arrange the steps in the correct order to share data between two Airflow tasks using XComs.

A 1,3,4,2

B 3,1,4,2

C 1,4,3,2

D 3,4,1,2

💡 Hint

Pushing happens during Task A execution, which must complete before Task B runs and pulls.

Task A pushes the value (1), then completes (3). Task B runs after Task A (4) and pulls the value (2).

✅ Best Practice

expert

What is the recommended way to avoid XCom data size issues?

XComs have size limits. Which approach is best to handle large data sharing between tasks?

A Compress data and push compressed bytes in XCom

B Store large data in external storage and push only a reference via XCom

C Split large data into many small XCom pushes

D Increase Airflow database size limits to accommodate large XComs

💡 Hint

Think about scalability and Airflow best practices.

Best practice is to store large data outside Airflow (e.g., S3, database) and push only a pointer or filename via XCom to avoid database bloat and performance issues.