Validatorian

Validations are a method of ensuring that our databases only receive the type of information that is appropriate for each attribute. After all, we wouldn't want a surprise type of data to find it's way into our code and cause unexpected behavior. Fortunately, SQLAlchemy has a package that makes validations quick and easy!

Let's look at some simple examples. Imagine we have a simple model, Sandwich. Here we have already initialized our database and are importing it from a configuration file.

from config import db

class Sandwich(db.Model):
    __tablename__ = 'sandwiches'

    id = db.Column(db.Integer, primary_key=True)
    name = db.Column(db.String)
    description = db.Column(db.String)
    price = db.Column(db.Float)

If we want to add validations to any of these attributes, we'll need to import the validations package first.

from sqlalchemy.orm import validates

And then, write our function with the '@validates' decorator inside the model.

    @validates('name')
    def validate_name(self, key, value):
        if not value:
            raise ValueError('Name cannot be empty.')
        if type(value) != str:
            raise ValueError('Name must be a string.')
        return value

So what's going on here? Let's break it down. The @validates is a decorator that lets our ORM know to pass any values received with the key of 'name' through our validation function before adding them to the database. The value we return is what is finally given to our database. The "key" argument is the key that is being evaluated, in this case 'name', the value is the value of that key, so the actual name(hopefully in text) that we are trying to add. So here we are checking to make sure that the name attribute passed in is not empty, and that is in in fact a string. If not, we raise an error.

We can also run multiple attributes through the same decorator by adding them to it's arguments.

    @validates('name', 'description')
    def validate_text(self, key, value):
        if not value:
            raise ValueError(f'{key} cannot be empty.')
        if type(value) != str:
            raise ValueError(f'{key} must be a string.')
        return value

This function validates our name and description attributes, but we won't usually have the same validations to perform on different attributes. Depending on how different and how many validations we have we can do it a couple of different ways. We can run a separate validator for our other attributes, let's add a length validation for our description and price validations too:

    @validates('name')
    def validate_name(self, key, value):
        if not value:
            raise ValueError('Name cannot be empty.')
        if type(value) != str:
            raise ValueError('Name must be a string.')
        return value

    @validates('description')
        def validate_description(self, key, value):
        if not value:
            raise ValueError('Description cannot be empty.')
        if type(value) != str:
            raise ValueError('Description must be a string.')
        if not 10 <= len(value) <= 200:
                raise ValueError('Description must be between 10 and 200 characters.')
        return value

    @validates('price')
    def validate_price(self, key, value):
        if type(value) != float:
            raise ValueError('Price must be a float.')
        if not 1 <= value <= 15:
            raise ValueError('Price must be between 1 and 15')

Or, we can keep the same validator for both and use the key argument passed in to adjust which validations are run for each attribute.

    @validates('name', 'description', 'price')
    def validate(self, key, value):
        if key != 'price:
            if not value:
                raise ValueError(f'{key} cannot be empty.')
            if type(value) != str:
                raise ValueError(f'{key} must be string.')
            if key == 'description':
                if not 10 <= len(value) <= 200:
                    raise ValueError('Description must be between 10 and 200 characters.')
        else:
            if type(value) != float:
                raise ValueError(f'{key} must be a float.')
            if not 1 <= value <= 15:
                raise ValueError('Price must be between 1 and 15')
        return value

Hmm, this is a little messy, let's refactor into 2 separate validators.

    @validates('name', 'description')
    def validate_text(self, key, value):
       if not value:
            raise ValueError(f'{key} cannot be empty.')
        if type(value) != str:
            raise ValueError(f'{key} must be string.')
        if key == 'description':
            if not 10 <= len(value) <= 200:
                raise ValueError('Description must be between 10 and 200 characters.')

    @validates('price')
    def validate_price(self, key, value):
        if type(value) != float:
            raise ValueError('Price must be a float.')
        if not 1 <= value <= 15:
            raise ValueError('Price must be between 1 and 15')

That's better! Here's our completed model:

from sqlalchemy.orm import validates
from config import db

class Sandwich(db.Model):
    __tablename__ = 'sandwiches'

    id = db.Column(db.Integer, primary_key=True)
    name = db.Column(db.String)
    description = db.Column(db.String)
    price = db.Column(db.Float)

    @validates('name', 'description')
    def validate_text(self, key, value):
        if not value:
            raise ValueError(f'{key} cannot be empty.')
        if type(value) != str:
            raise ValueError(f'{key} must be string.')
        if key == 'description':
            if not 10 <= len(value) <= 200:
                raise ValueError('Description must be between 10 and 200 characters.')

    @validates('price')
    def validate_price(self, key, value):
        if type(value) != float:
            raise ValueError('Price must be a float.')
        if not 1 <= value <= 15:
            raise ValueError('Price must be between 1 and 15')

That's it! Validations are one easy tool to make sure your databases stay right and proper.