1

So, I have a table in mysql that stores user interests, initially it was just data to analyze when needed. The users_interests table is a table that stores the user_id, the section_id, the location_id, the property_type_id, the minimum price and the maximum price. This table now contains about 6 million records, and now we need to analyze it in real time, but mysql is not helping (very slow and sometimes timeout).... What I need to achieve is building a form that has many filters (that are optional) with which the user can filter (Every time the user chooses a filter it will send an api to the database to count the unique number of users that have interest with that criteria).

The form should contain fields for section_id, property_type_id, location_id, min_price, max_price and country code.... And each time the user chooses a filter for example section_id = 1, it should send an api to count unique users in real time that have interests in section_id = 1 and respond in less than a second optimally, then when the user select property_type_id = 4 then we will send another api to count unique users in real time that have interest in section_id = 1 and property_type_id = 4 and etc as the user keeps filtering.

As I mentioned, I tried with mysql and it was really slow and sometimes would timeout (even with indexing, I tried also composite indexing but they won't work as all of my filters are optional) ... I read into this problem, and a lot of people mentioned the hyperloglog algorithm, which what I gathered approximates the distinct count to optimize performance. So are there big data solutions that support the hyperloglog algorithm? And are they sufficient to help me solve my problem?

0

Have you looked into big data solutions such as Hive/Hadoop for storage, Spark for stream processing, and Kafka for messaging? You could use an existing solution such as Google's Big Table.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.