Flask API请求体验证：允许useCaseId字段支持字符串与字典类型

Got it, let's tweak your Marshmallow schema so useCaseId accepts both strings and dictionaries—here are two solid ways to do it:

Method 1: Custom Validator with fields.Raw (Works for all Marshmallow versions) #

This approach uses a custom validation function to check that the value is either a non-empty string or a dictionary, and works with any Marshmallow version.

from marshmallow import Schema, fields, validate, ValidationError

# This function validates the request body
def validatefilterdata(data):
    # Custom validator to enforce useCaseId is either string (min length 1) or dict
    def validate_use_case_id(value):
        if isinstance(value, str):
            if len(value) < 1:
                raise ValidationError("useCaseId string must be at least 1 character long")
        elif not isinstance(value, dict):
            raise ValidationError("useCaseId must be either a string or a dictionary")

    class FilterDataSchema(Schema):
        masterBotId = fields.Str(validate=validate.Length(min=1), required=False)
        # Updated field to use Raw + custom validator
        useCaseId = fields.Raw(validate=validate_use_case_id, required=False)
        month = fields.Str(required=False)
        dateRange = fields.Dict(required=False)
    
    filetdata_schema = FilterDataSchema(strict=True)
    validation_result = filetdata_schema.load(data)
    return validation_result.errors

How it works: #

• fields.Raw is the base class for all Marshmallow fields, so it doesn't enforce a specific data type by default.
• Our custom validate_use_case_id function checks if the value is a string (and validates its length) OR a dictionary. If neither, it throws a clear validation error.

Method 2: Use fields.Union (Marshmallow 3.0+ only) #

If you're running Marshmallow 3.0 or newer, the Union field is a cleaner, more idiomatic solution. It lets you define multiple acceptable field types, and Marshmallow will try each one until it finds a match.

from marshmallow import Schema, fields, validate, ValidationError

def validatefilterdata(data):
    class FilterDataSchema(Schema):
        masterBotId = fields.Str(validate=validate.Length(min=1), required=False)
        # Union field accepts either a non-empty string or a dictionary
        useCaseId = fields.Union([
            fields.Str(validate=validate.Length(min=1)),
            fields.Dict()
        ], required=False)
        month = fields.Str(required=False)
        dateRange = fields.Dict(required=False)
    
    filetdata_schema = FilterDataSchema(strict=True)
    validation_result = filetdata_schema.load(data)
    return validation_result.errors

How it works: #

• fields.Union takes a list of field types. When deserializing, it will attempt to validate the value against each field in order.
• If the value is a string, it passes through the Str field's length validation. If it's a dictionary, it passes through the Dict field. Any other type will trigger a validation error.

Pro tip: If you need to add more context to the error messages for Union, you can pass an error_messages parameter to the Union field to make errors clearer for API consumers.

内容的提问来源于stack exchange，提问作者Tony Mathew