Django DRF many=True时未调用ListSerializer的validate方法问题排查

问题根因 #

DRF 中many=True模式下的序列化器校验流程是固定的，执行顺序如下：

1. 首先校验传入数据为合法列表格式
2. 遍历列表中每一个元素，先执行子序列化器的单字段校验，再调用子序列化器的validate()方法完成单条数据的业务校验
3. 所有单条数据全部校验通过后，才会调用ListSerializer子类的validate()方法，传入所有已校验的单条数据做列表级全局校验

你代码的问题在于：子序列化器RecipeStepDraftCreateSerializer的validate()方法里写入了仅单条提交场景才需要的校验逻辑（比如step_number必须等于1、必须等于现有步骤最后序号+1），批量提交时第二条及之后的元素在经过子序列化器校验时就会触发报错，根本不会走到你写的批量序列化器校验步骤。

修复方案 #

只需调整子序列化器的validate()逻辑，将通用校验（如字段非空）和单条提交专属校验拆分，批量模式下跳过单条专属校验即可：

修改后的 serializers.py 代码 #

class RecipeStepDraftBulkCreateSerializer(serializers.ListSerializer):
    # 原有逻辑无需修改
    def validate(self, data):
        print("bulk validate")
        new_step_numbers = [s.get('step_number') for s in data]
        if None in new_step_numbers:
            raise serializers.ValidationError("step_number filed is required")
        if new_step_numbers != list(set(new_step_numbers)):
            raise serializers.ValidationError("Wrong order of step_number(s) supplied")
        try:
            recipe = Recipe.objects.get(pk=self.context.get('recipe_id'))
            existing_steps = recipe.recipe_steps.get_queryset().all()
            if existing_steps:
                ex_step_numbers = [s.step_number for s in existing_steps]
                if new_step_numbers[0] != ex_step_numbers[-1] + 1:
                   raise serializers.ValidationError(
                    f"The next first supplied step_number must be: {ex_step_numbers[-1] + 1}") 
                steps_combined = ex_step_numbers + new_step_numbers
                if steps_combined != list(set(steps_combined)): 
                    raise serializers.ValidationError(f"Wrong order of step_number(s) supplied")  
            return data   
        except ObjectDoesNotExist:
            raise serializers.ValidationError("Recipe under provided id doesn't exist.")    
    def create(self, validated_data):
        recipe = Recipe.objects.get(pk=self.context.get('recipe_id'))
        for step in validated_data:
            step['recipe'] = recipe
            RecipeStep.objects.create(**step)
        return validated_data                

class RecipeStepDraftCreateSerializer(serializers.ModelSerializer):  
    class Meta:
        model = RecipeStep 
        fields = [ 
            'id',
            'step_number',
            'step_image',
            'instruction',
            'tip']  
        list_serializer_class = RecipeStepDraftBulkCreateSerializer
    
    def validate(self, data):
        print("single validate")
        # 通用校验：单条、批量场景都需要执行
        if not data.get("step_number"):
            raise serializers.ValidationError("step_number field is required.")
        try:
            recipe = Recipe.objects.get(pk=self.context.get('recipe_id'))
            # 仅非批量场景（没有父级ListSerializer）执行单条提交专属校验
            if not isinstance(self.parent, RecipeStepDraftBulkCreateSerializer):
                existing_steps = recipe.recipe_steps.get_queryset().all()
                if existing_steps:
                    ex_step_numbers = [s.step_number for s in existing_steps]
                    if data["step_number"] != ex_step_numbers[-1] + 1:
                        raise serializers.ValidationError(
                        f"The next first supplied step_number must be: {ex_step_numbers[-1] + 1}") 
                if data["step_number"] != 1:
                    raise serializers.ValidationError(f"Wrong step_number. 1 expected, got {data['step_number']}")        
            return data
        except ObjectDoesNotExist:
            raise serializers.ValidationError("Recipe under provided id doesn't exist.")
    def create(self, validated_data):
        # 原有逻辑无需修改
        recipe = Recipe.objects.get(pk=self.context.get('recipe_id'))
        validated_data['recipe'] = recipe
        step = RecipeStep.objects.create(**validated_data)
        return step 

逻辑说明 #

• 子序列化器通过self.parent判断当前是否被批量序列化器调用，批量场景下跳过单条专属的序号校验逻辑，只做必填项等通用校验
• 列表级的序号连续、不重复、起始值正确等逻辑完全由批量序列化器的validate()方法负责，符合你最初的业务拆分预期
• views.py 原有逻辑无需任何调整

内容的提问来源于stack exchange，提问作者Anton Savitskiy