Hacker News Reader

Comment by gjmacd | Hacker News Reader

Someone needs to explain to me what the benefits of NoSQL with MongoDB are when you have the JSONB column type and the ability to query and insert at the field level with JSON in PostgreSQL? Maybe there's some benefit, but I'm not seeing that major "gotta have it" feature or performance gains. And I ask this question seriously, because I just don't know the answer.

View on HN

>Someone needs to explain to me what the benefits of NoSQL with MongoDB are when you have the JSONB column type

Schema-less design is a feature of MongoDB, not NoSQL, but has unfortunately has persisted as one of the "benefits" of NoSQL. However when you look at the top 5 "NoSQL" databases on db-engine rankings only 2 support "JSON", MongoDB and Elasticsearch. The rest are key value, or require schemas (Cassandra).

The appeal of NoSQL, to me, has been the scale out to tens, hundreds or thousands of machines relatively easily. This is where Postgres does not shine, and this is the only reason I'd choose something as binding as DynanmoDB over something like Postgres.

Having to work with MongoDB, I say only benefit is you don't need to plan or think through your design. Which in my view is not really a benefit. You end up checking for null properties everywhere.

Most of the positives here seem to be "it's easy". Someone said they had to manage their indexes on it, which was bad.

We can't just keep relying on super fast hardware and magical software to get us out of having to think.

Reply on news.ycombinator.com