Rushing Labs

Learning Python CRUD - Initial Choices & Issues

Learning a new thing§

Upon starting a new job, I was challenged by for the first time being in a place that doesn't use any language I have previous/extensive experience in. And we have quite the variety too: mostly Python, Java, JavaScript...and sparse bash, Powershell, PHP, PL-SQL, and a few others.

But! our infosec team has chosen to standardize on Python for any new tooling that needs writing, so here's to me diving into Python by trying to build a small web app. Ideally, I'd like to build a little budgeting app (the UI similar to EveryDollar), but with a few extra features for pulling in email receipts and balancing records from my bank.

So, I thought it would be interesting to capture some first impressions about getting into Python and these frameworks. Learning curve and mistakes imminent...

Flask vs FastAPI§

  • Flask seems biased towards monolithic apps

    • it's possible to create REST API/data-only service-based apps, but it feels like you aren't supposed to.
  • FastAPi has a bend toward asynchronous, service-based APIs

  • Not sure if this is actually how either of these frameworks are structured, but that is the "feel" I get after working with both of them.

UUID issue§

Describing UUIDs as UUID(as_uuid = True) was necessary in sqlalchemy models.

 id = Column(UUID(as_uuid=True), primary_key = True, default=uuid.uuid4) 

This piece stumped me for a bit. Specifically because, to me, Column(UUID) looks like it should describe the data type for this column just fine. Primary key is fine. Setting a default for the server is fine. But why specify a UUID is going to be written as_uuid=True...what else would it be?? I'm sure there's a good reason, but it stumped me for a bit, lol

Direct vs. Declarative for sqlalchemy Models§

The one thing I would have to say is most difficult about learning Python is the disparate documentation. Documentation certainly exists, but it's not always consistent. Admittedly, I think I've been spoiled by the C#/.NET ecosystem that I started programming in. There isn't the same type of relationship between tools and docs in Python as there is with Visual Studio + Intellisense + Microsoft Docs + .NET libaries & APIs. Not a detriment to Python, but increased the learning curve for me.

Such, as running into the issue where I needed to know how to declare relationships within my sqlalchemy models:

Then, I find out the example in that link (directly above) was using the "declarative" Table class, and some other tutorial/article guided me towards using a "direct" Table class:

Debugging FastAPI§


More on that topic of tools + language ecosystem integration, debugging always seems to be a questionable area. Debugging a simple Python script couldn't be easier: (1) load in VS Code, (2) set breakpoints, (3) run. However, using a framework like Flask, Fast API, or anything else seems to mean finding a debug configuration, and loading that into your editor.

So here's the configuration needed for FastAPI, in launch.json, for VSCode. The extra pieces in the args section are important. Throw this in your launch.json, and you should be good to go.

"version": "0.2.0",
  "configurations": [
      "name": "Python: FastAPI",
      "type": "python",
      "request": "launch",
      "module": "uvicorn",
      "cwd": "${workspaceFolder}/<folder to your>",
      "args": [
        "--port", //these arg are optional


Overall, having used Python sparingly before diving into these web frameworks, it definitely required less effort than I expected to get something functional on screen. That aspect of Python continues to amaze me. Whether it's web frameworks, AI/ML tooling, or scripts, Python always seems to require minimal developer-hand-written code to produce what you want on screen. But to build a full-featured, robust, secure API-driven web application...I don't know if I would want to commit to doing so in Python. The feeling of a full ecosystem, tooling, docs, and support just doesn't seem to be there.