Django: Use MySQL and ElasticSearch in single application

Question

I want to use MySQL and ElasticSearch using Haystack within single Django application. Our's application deals with a huge set of relational data with 300+ tables and MySql is the best choice to serve that. But we want fast searching on few columns of tables with 10 Million+ rows.

Is it possible in Django to use MySQL as primary data storage and use ElasticSearch for search related queries? I searched on internet but couldn't find any related information.

• If yes, what should be the approach? Any related links or blogs will be very helpful.
• If not, what should be the alternative approach?

Answer 1

Yes, we can configure MySQL and ElasticSearch within a single Django project. Firstly install django-haystack and pyelasticsearch packages from pip.

sudo pip install django-haystack
sudo pip install pyelasticsearch 

In settings.py, add 'haystack' to INSTALLED_APPS and configure HAYSTACK_CONNECTIONS like:

HAYSTACK_CONNECTIONS = {
    'default': {
        'ENGINE': 'haystack.backends.elasticsearch_backend.ElasticsearchSearchEngine',
        'URL': 'http://<ip-of-redis-server>:9200/',
        'TIMEOUT': 60 * 5,
        'INCLUDE_SPELLING':True,
        'INDEX_NAME': '<name-of-index>'
        }
    }

Create a class inheriting hystack.indexes.SearchIndex and hystack.indexes.Indexable classes. In the below sample code, index in created on the fields of MyModelClass model:

from haystack import indexes


class MyIndexClass(indexes.SearchIndex, indexes.Indexable):
    field_1 = indexes.CharField(document=True, use_template=True)
    field_2 = indexes.CharField(model_attr='field_of_MyModelClass_class')
    field_3 = indexes.CharField(model_attr='field_of_MyModelClass_class')

def get_model(self):
    return MyModelClass

Refer HayStack document for further details.