如何在Cerberus Schema中结合excludes与default实现字段校验归一化？

Alright, let's tackle your Cerberus schema problem head-on. You need two key behaviors: block documents that have both value_1 and value_2, and auto-set value_1 to 0 when it's missing. Plus, you want to understand how to combine excludes and default rules properly. Here's the breakdown:

Step 1: Build the Schema with Combined Rules #

First, let's write the schema that meets both requirements. We'll use excludes to enforce mutual exclusivity and default to populate value_1 when it's absent:

from cerberus import Validator

# Define the schema
schema = {
    'value_1': {
        'type': 'integer',
        'default': 0,  # Auto-set to 0 if missing
        'excludes': 'value_2'  # Can't exist if value_2 is present
    },
    'value_2': {
        'type': 'integer',
        'excludes': 'value_1'  # Can't exist if value_1 is present
    }
}

Step 2: How It Works #

Let's break down each rule's role:

• excludes Rule: This rule ensures that if one field exists in the input document, the other cannot. We add it to both fields to make the exclusivity bidirectional—so whether value_1 or value_2 is present first, the other gets blocked.
• default Rule: Applied to value_1, this fills in the value 0 automatically only when value_1 is not present in the input. Since we're not marking value_1 as required, this only triggers on missing fields.

Step 3: Test the Schema #

Let's verify it works for your target scenarios:

Scenario 1: Input has only value_2 (your desired normalization case) #

v = Validator(schema)
input_doc = {'value_2': 1}

if v.validate(input_doc):
    print(v.normalized(input_doc))  # Output: {'value_1': 0, 'value_2': 1}
else:
    print(v.errors)

This gives exactly the normalized result you want—value_1 is auto-set to 0, and no validation errors are thrown.

Scenario 2: Input has both value_1 and value_2 (should trigger error) #

input_doc = {'value_1': 5, 'value_2': 1}

if v.validate(input_doc):
    print(v.normalized(input_doc))
else:
    print(v.errors)  # Output: {'value_1': ['cannot be present with value_2'], 'value_2': ['cannot be present with value_1']}

As expected, Cerberus throws a validation error telling you the two fields can't coexist.

Key Notes on Combining excludes and default #

• Order of Operations: Cerberus first checks validation rules (like excludes) against the original input, then applies default values. This means filling value_1 via default won't trigger the excludes rule, because the original input didn't have value_1—only the normalized output does, and validation already passed.
• Bidirectional Exclusivity: Don't forget to add excludes to both fields. If you only added it to value_1, the schema would block value_2 when value_1 exists, but allow value_1 to be added via default even if value_2 is present (which isn't what you want).
• Type Safety: We added type: 'integer' to both fields to ensure inputs are valid numbers—adjust this if your use case allows other types.

内容的提问来源于stack exchange，提问作者El Ruso