SQL can now replace Python for many supervised ML tasks. Must you make the switch?
Relating to machine learning, I’m an avid fan of attacking data where it lives. 90%+ of the time, that’s going to be a relational database, assuming we’re talking about supervised machine learning.
Python is amazing, but pulling dozens of GB of knowledge at any time when you desire to train a model is a big bottleneck, especially if you might want to retrain them incessantly. Eliminating data movement makes quite a lot of sense. SQL is your friend.
For this text, I’ll use an always-free Oracle Database 21c provisioned on Oracle Cloud. I’m undecided for those who can translate the logic to other database vendors. Oracle works like a charm, and the database you provision won’t cost you a dime — ever.
I’ll leave the Python vs. Oracle for machine learning on huge dataset comparison for another time. Today, it’s all about getting back to basics.
I’ll use the next dataset today:
So download it to follow along, and be certain that you’ve a connection established to your database instance. Tools like SQL Developer or Visual Studio Code can try this for you.
Find out how to Create a Database Table
The next code snippet will create the iris table. The id column is mandatory, as Oracle will need it behind the scenes:
create sequence seq_iris;create table iris(
id number default seq_iris.nextval,
sepal_length number(2, 1),
sepal_width number(2, 1),
petal_length number(2, 1),
petal_width number(2, 1),
species varchar2(15)
);
Once the table is created, it’s time to load the info.