Airflow DAG: Coding your first DAG for Beginners
ฝัง
- เผยแพร่เมื่อ 27 ก.ย. 2024
- Airflow DAG, coding your first DAG for Beginners.
👍 Smash the like button to become an Airflow Super Hero!
❤️ Subscribe to my channel to become a master of Airflow
🏆 BECOME A PRO: www.udemy.com/...
🚨 My Patreon: / marclamberti to support my work and be a friend for life
Starting with Apache Airflow can be difficult.
What is a DAG? What is an Operator? How DAGs are scheduled? so many questions. Well, you've come to the right place!
In this video, you will discover how to code your first DAG, the core concepts to understand and how to schedule your DAG.
Ready? Go!
The Code
www.notion.so/...
How to run Airflow locally with Docker
• Running Airflow 2.0 wi...
All you need about XComs:
marclamberti.c...
Url to the blog post:
marclamberti.c...
Thank you all for your warm feedback ❤ Here is another video to create a more advanced pipeline with AWS and Snowflake:
th-cam.com/video/wT67h9qDl1o/w-d-xo.html
Enjoy ❤
I can't express how grateful I am to you for sharing this content here with us on youtube.
Thank you and keep doing this excellent job.
Glad you enjoy it! :)
amazing explanation of the first DAG creation in airflow! Thanks a lot
Superb Narration about Airflow, with one video and simple example you cleared all my basic doubts. Thanks a lot.
Glad it was helpful!
You are a killer instructor! Following your tutorials feels like drinking French vanilla. Thumbs up!
this is very clear and insightful for me as a beginner, thank you! Can't wait to try it on my own
Thank you 🙏
Thank you for sharing! I learned something new today! I appreciate your time!
Happy to help
Merci beaucoup Marc, bon courage.
Thank you sir, i really enjoyed learning while watching your video. Its the first time I discover your channel, definitely I'll be sharing it with my colleagues
Thanks that was amazing explanation
You’re welcome ❤️
Great explanation! I still wonder how the PythonOperator would be able to make an instance of a python class and call a specific method of that class. Most of the videos I have found only seem to showcase the use of functions for the python_callable param. 🤔
best tutorial on airflow DAG ✌
thanks a lot! it really help me get going with dags
Happy to help
That explanation is really good. Kudos!
The best explanation, kudos to you
That was both informative and enjoyable. Thank you Marc!
It will be great if you include in the tutorial how to open a file, save it and run it using airflow.
Thanks Marc. Very well explained.
Thank you vry much) I an working with my homework with your video
Wonderful explanation. Thank you very much for the video!
This was incredible.. Thank you Mark
u are amazing man. so clear !
Very comprehensible. Thank you!
Really helpful! Thanks from Québec!
In Function, _choose_best_model return "accurate" .
How does the python/airflow know that "accurate" is not a string but a task_id for BashOperator ?
It won't be python that "knows", it will be what airflow is doing behind the scenes. Since he's specified it as a python_callable, I imagine airflow will call the function and respond to it's return appropriately.
Thank you! I started to understand...
awesome explanation!
Thank you. All simply and helpful
Awesome, man. Many thanks!
I have written code into Jupyter notebook it successfully executed over here ...
Very useful ! Thank you for the sharing!
really good content, thanks Marc!
very well explained.. thanks
brilliant and simple!
Awesome explaination!!
Amazing work
I'm new with airflow..... currently I have a server with jupyterhub+jupyterlab...I've installed airflow at the same server and I wanted to create this DAG from jupyterlab... but I don't have visibility of airflow modules within jupyter environmente despite of they are installed at the same server... How can i proceed?... and leads me to this question, where should I build one dag? what's your suggestion?
Very good tutorial
Great teaching skill. Thank you for the tut
This is great, thank you!
happy to help! :)
I have airflow up and running. but it is unable to import airflow library. Any help
Thanks brother!
thank you
Awesome explanation
Glad you liked it
How to implement the condition where accurate should run only when training model A,B,C all 3 are successful executed?
Great video! TY!
How to call all snowflake stored procedures with one Task in another Python file , when corresponding Operators in declared in Main DAG File
How can we put best_accuracy on output?
How did you submit your script to Airflow? Only then you'll be able to view it in Web UI right?
Sorry, I did not find any video in description that explain how to install Airflow to my PC. Can you help me, please ?
Really great content!
I think someone already asked. Do you also need to install apache-airflow locally with pip in order to get code completion? Thanks for the great content!
Yes
i have a question! How i can see result of pipeline. For example i have a function print('hello world') and i want to see it in screen
Hello sir ,I have created dags successfully but it is not visible at airflow web interface what should we have to do ?
Hi Marc, please suggest me your Udemy coarse, as I am working in GCP composer
Thank you for the great content
Brilliant!
I didn't find the link in description
Marc, I stuck with an issue. I am trying to create multiple dagRun with same execution time, but getting exception. To overcome this, i tried to create it with microsecond precision, but still dagRuns are using "seconds" and truncating the microseconds. I also tried "replace_microseconds"=false, but no success. Please help or if you know any doc, please share.
azure data factory with azure databricks is awesome. Airflow is cave man tool.
Hi - I have passed this JSON {"Name" : "Jhonny"} in configuration JSON box before triggering manually. I want to print last two letters of the value which passed to the Name i.e. in this example "ny"..How do I print this in Airflow DAG..I am unable to print it
Hello, thanks for the content, but some probleme, when i run the dag , hava a error ERROR - name 'best_accuracy' is not defined
Nice instructor
Hi @marclamberti
I want ask as a Data Engineer, I want to regularly clean up airflow log file that more than 2 months old. Is it possible?
I can see the dag in the airflow UI but it never runs for me.
Marc, I reproduced the example you demonstrated, but I notice strange behavior: when the function fetches results from the training runs, the results are the same each time I run the DAG, so the same branch is always taken. It seems like the training function result gets cached and re-used. Any idea why?
Update: I added a call to random.seed() in the _training_model function, and it resolved the problem.
great content
Wonderful 👏 👏 👏
Great video!
Thank you Raul 😁
mssqloperator and mssqlhook airflow example pls
Marc you are incredibly good at explaining. Perfect balance between details and conciseness! Finished this exercise succesfully at the first try! One thing I still do not understand is how can I have a task launch some external python programs that are managed in their own virtual environments by Poetry? Thanks
Hi Marc, Awesome lecture. Though I have a small doubt. Lets say I am currently working on Azure cloud. I am using databricks jobs for my ETL. Then why should I learn airflow if I can schedule my job dependencies using Azure data factory? What are the advantages over other data integration tool? I am confused about this one thing.
It's not only about ADF or airflow, there are hundereds of scheduling/orchestration tools out there. You need to see which one works for you. Your question can be rephrased as we already have IBM cloud and AWS, why do we need Azure. The simple thing to understand is that every tool/service provides features, you need to cehck which one works for you. One way to go is, do some research, read few articles. What I would recommend is, read about few tools, choose 2 best tools based on features they provide, take 5 days, work on 2 POCs around your use case, weigh the pros and cons, you should have better understanding. There can be other factors depending upon the company/institute you are at, if you require good prompt support, the associated cost etc. Go for the research, try POCs and make an informed decision. Don't be afraid to make mistakes, that's how we all learn.
I forgot to tag the link in the above message(PS:I have no idea about ETL or ADF)
www.elixirdata.com/blog/azure-data-factory-vs-apache-airflow#:~:text=Azure%20Data%20Factory%3A%20It%20supports,directed%20acyclic%20graphs%20of%20tasks.
Will training model A,B,C be executed in parallel ?
Yes
@Marc Lamberti, to answer your question. To reduce the repeated line of code for Training_Model_A, B and C. we can create python function for the same and call it 3 times. Please let us know your thoughts. Thanks for good content. Appreciate it.
Does it work? Have you tried?
Excellent tutorial! Just one question: is there any particular reason to use functions with an underscore, like "_training_model" instead of just "training_model"?
"_" is usually used to indicate that this function belongs to same file.
It's an indication that this is an private function or for internal use only.
How did you actually deploy that code into Airflow? I think that's a very important step skipped in this tutorial. Without that this tutorial is just how to write python code for Airflow. :(
Well yo deploy that code in Airflow you just add it into the dags folder and that's it :)
i think using loop we can avoid repeat the task again and again.
i thought DAGs are acyclic
Cool example pipeline for ML (instead of boring ETL)
Thank you for the great video! Is the midnight of the datetime that it starts to run the UTC time or the local time?
Utc
I cant import airflow on Visualstudio... how can I make
Hi, Fael! I’m having the same issue. Did you solve the problem?
Hi bro u looks like Johny sinns😂😂
Sir, is there any way to get list of dag id using python ?
Thomas Laura Williams Cynthia Walker Helen
Not a good example, its all over the place!
why isn't a good example?
@@MarcLamberti He doesn't know what he is talking about. I'm a total newbie and followed along perfectly :)
I was looking for help debugging the "./start.sh" in airflow-section-4 in the Udemy course "Apache Airflow - The Hands On Guide" Error Preparing metadata (setup.py): started
I purchase the course in Udemy
Great Tutorial
very good tutorial
how do you run locally the airflow UI? when I use airflow standalone command it tells me: 'airflow airflow Invalid login. Please try again.'
How can I integrate those Deep learning model into spark or airflow, can you make a video about this like how we can integrate our ML or DL model into Airflow or spark for job scheduling
Great video! So helpful! Do a video on ETL airflow but loading into postgres or with sql operators
The PostgresOperator is the way 😁
Clear explanation for the beginners. Thank you!
Nacho Vargo teach airflow
love it, great video to start getting hands on airflow! please keep making more videos like these using different and more complex scenarios.
this playlist not arranged for beginners and intermediate plesae if you can arrange it or make it 2 playlists
Oh! That's good idea! Will do that, thanks
Use list comprehension - e.g.:
training_models = [ PythonOperator(task_id=f"training_model_{step}", python_callable=_training_model) for step in ['A','B','C'] ]
amazing one-liner. Only improvement I can think of is to specify the model list globally so it can be specified in the task creation as well as the xcom call
How you are able to get suggestions in your VSC without installing the Airflow dependencies?
This is very informative
how do i import a json config file that stores variables in another python script with airflow
I am running airflow on port 8002. How to get my_dag in the panel?
Hey! Thanks for great videos. I am facing trouble while running a java jar file from airflow. Getting java command not found error message.
P.s- tried with adding path in $PATH. Can not use docker.
but,,, how my_dag appears in monitor ? by magic ?