有没有办法在序列化数据之后只获取查询集中的特定字段，而不在 drf 中创建不同的序列化程序？

Question

i need to do a query where i want to get specific fields, then serializate it and keep only the specific fields which I got in the query.

models.py

class Search(models.Model):
    NEUTRAL = 'None'
    POSITIVE = 'P'
    NEGATIVE = 'N'
    POLARITY_CHOICES = [
        (NEUTRAL, 'Neutral'),
        (POSITIVE, 'Positive'),
        (NEGATIVE, 'Negative'),
    ]
    user = models.ForeignKey(User,related_name='searched_user_id',on_delete=models.CASCADE)
    word = models.CharField( max_length = 100)
    social_network = models.ForeignKey(SocialNetwork,related_name='search_social_network_id',on_delete=models.CASCADE)
    polarity = models.CharField(
        max_length=4,
        choices=POLARITY_CHOICES,
        default=NEUTRAL,
    )
    sentiment_analysis_percentage = models.FloatField(default=0)
    topic = models.ForeignKey(Topic,related_name='search_topic_id',on_delete=models.CASCADE)
    liked = models.IntegerField(default=0)  
    shared = models.IntegerField(default=0) 
    is_active = models.BooleanField(default=True)
    is_deleted = models.BooleanField(default=False)
    updated_date=models.DateTimeField(auto_now=True)
    searched_date = models.DateTimeField(auto_now_add=True)

serializers.py

class SearchSerializer(serializers.ModelSerializer):
    searched_date = serializers.DateTimeField(format="%d-%m-%Y")
    class Meta:
        model = Search
        fields = ('__all__')

class RecentSearchSerializer(serializers.ModelSerializer):
    searched_date = serializers.DateTimeField(format="%d-%m-%Y")
    class Meta:
        model = Search
        fields = ('user','social_network','word','searched_date')

class SentimentAnalysisSerializer(serializers.ModelSerializer):
    searched_date = serializers.DateTimeField(format="%d-%m-%Y")
    class Meta:
        model = Search
        fields = ('polarity','searched_date','sentiment_analysis_percentage')

SearchSerializer is the main serializer for search, RecentSearchSerializer is the serializer to pass data and filtering in the DRF api view, and finally I created SentimentAnalysisSerializer to keep the specific fields that I need:

api.py

class SearchViewSet(viewsets.ModelViewSet):
    queryset = Search.objects.filter(
        is_active=True,
        is_deleted=False
    ).order_by('id')
    permission_classes = [
        permissions.AllowAny
    ]
    serializer_class = SearchSerializer
    pagination_class = StandardResultsSetPagination

    def __init__(self,*args, **kwargs):
        self.response_data = {'error': [], 'data': {}}
        self.code = 0

    def get_serializer_class(self):
        if self.action in ['recent_search','word_details']:
            return RecentSearchSerializer
        return SearchSerializer

    @action(methods=['post'], detail=False)
    def word_details(self, request, *args, **kwargs):
        try:
            self.response_data['data']['word'] = kwargs['data']['word']
            queryset = Search.objects.filter(
                is_active=True,
                is_deleted=False,
                social_network=kwargs['data']['social_network'],
                user_id=kwargs['data']['user'],
                word=kwargs['data']['word']
            ).order_by('id')
            import pdb;pdb.set_trace()
            serializer = SentimentAnalysisSerializer(queryset, many=True)
            self.response_data['data']['timeline_word_twitter_polarity'] = json.loads(json.dumps(serializer.data))

I did this solution and works good, but is there a way to have the same behaviour without create another serializer? I mean, using SearchSerializer?

I tried with the following examples and i got these erros:

(Pdb) queryset = Search.objects.filter(is_active=True,is_deleted=False,social_network=kwargs['data']['social_network'],user_id=kwargs['data']['user'],word=kwargs['data']['word']).values('polarity','sentiment_analysis_percentage','searched_date').order_by('id')
(Pdb) serializer = RecentSearchSerializer(queryset, many=True)
(Pdb) self.response_data['data']['timeline_word_twitter_polarity'] = json.loads(json.dumps(serializer.data))
*** KeyError: "Got KeyError when attempting to get a value for field `user` on serializer `RecentSearchSerializer`.\nThe serializer field might be named incorrectly and not match any attribute or key on the `dict` instance.\nOriginal exception text was: 'user'."
(Pdb) 

(Pdb) queryset = Search.objects.filter(is_active=True,is_deleted=False,social_network=kwargs['data']['social_network'],user_id=kwargs['data']['user'],word=kwargs['data']['word']).values('polarity','sentiment_analysis_percentage','searched_date').order_by('id')
(Pdb) serializer = SearchSerializer(queryset, many=True)
(Pdb) self.response_data['data']['timeline_word_twitter_polarity'] = json.loads(json.dumps(serializer.data))
*** KeyError: "Got KeyError when attempting to get a value for field `word` on serializer `SearchSerializer`.\nThe serializer field might be named incorrectly and not match any attribute or key on the `dict` instance.\nOriginal exception text was: 'word'."

First I thought that those errors were related with this issue , but according with this answer i'm not passing data parameter like the issue explain, so i can't check what is the error with the Validation method (is_valid())

I'm using the last version of DRF: djangorestframework==3.10.3

I wish to get this result but with SearchSerializer (I need to do other queries with specific fields, i mean I don't need to pass al the fields of Search Model), but I don't know if it's possible

(Pdb) serializer = SentimentAnalysisSerializer(queryset, many=True)
(Pdb) self.response_data['data']['timeline_word_twitter_polarity'] = json.loads(json.dumps(serializer.data))
(Pdb) self.response_data['data']['timeline_word_twitter_polarity']
[{'searched_date': '09-10-2019', 'polarity': 'P', 'sentiment_analysis_percentage': 0.0}, {'searched_date': '09-10-2019', 'polarity': 'N', 'sentiment_analysis_percentage': 0.0}]

Thanks in advance, any help will be appreciated.

Answer 1

Well, the errors are clear.

You limit the query to return only certain fields using values . So then the serializer cannot serialize it because many are missing.

However, the following approach should work for you.

Note: I am not fan of this - i would rather have 2 separate serializers like you do. But it might help you.

class SearchSerializer(serializers.ModelSerializer):
    searched_date = serializers.DateTimeField(format="%d-%m-%Y")
    class Meta:
        model = Search
        fields = ('__all__')

    def __init__(self, instance=None, data=empty, **kwargs):
        super(SearchSerializer, self).__init__(instance, data, **kwargs)
        if instance is not None and instance._fields is not None:     
            allowed = set(instance._fields)
            existing = set(self.fields.keys())
            for fn in existing - allowed:
                self.fields.pop(fn)

Basically, it keeps only fields from the provided instance .