7 ms·

Someone please help me understand. I use relational databases everyday for my projects (db, api, frontend). I'm well-versed in SQL and general best practices, a

Someone please help me understand. I use relational databases everyday for my projects (db, api, frontend). I'm well-versed in SQL and general best practices, and my projects scale well (admittedly I'm not in the petabytes level of scale). Does the desire to use NoSQL databases come from a disdain for having to think in terms of declarative programming instead of imperative? Is it desirable to have dynamically typed fields in a database? I've always thought that the biggest motivation to use NoSQL is simply due to high-level language programmers' unfamiliarity with relational database logic and SQL syntax. Once I grasped the concepts, the relational model has proven indispensable for me. I will say, though, that it is important to get your schema correct at the beginning. I've also always worked full-stack, so I don't know what it's like not to have the ability to manipulate the database.

wild_preference 8y ago
NoSQL is used by two camps. One wants a denormalized doc store for their technical needs, and the other because it has an easier quickstart on day 1 when you don’t know SQL. Here we are usually talking of the latter. The beauty of something like Postgres is that you don’t need to get your schema right on day 1. How could you anyways unless tou already know all future data access pattern?
kodablah 8y ago
For me, it's two main reasons: 1) Generating and storing massive amounts of pre-computed permutations of the same data based on how I'll read it with easy linear growth by adding servers and 2) Transparent multi-master reads/writes providing easy HA failover. Note, not all NoSQL options offer this. If existing free relational DBs had these options easy and out of the box, I would use them instead. I just need a ton of data, often redundant, that I never delete, that supports distributed writes and reads (ACID not required). The things I use Cassandra for would require huge Postgres servers and custom sharding/partitioning setups and Citus HA and all sorts of other things.
njharman 8y ago
> what it's like not to have the ability to manipulate the database. In general "not wanting to understand, depend on, or work with other systems/teams" (often cause of dysfunctional eng/pm organizations) has, over 20years experience, been the number one reason devs (esp front end devs cause they tend to move the fastest (faster than traditional backend teams are prepared for) and have least experience in back end systems) go for something new, something they can control, something they can spin up fast and quick. NoSQL, cloud, anything they can do themselves.

wild_preference 8y ago

NoSQL is used by two camps. One wants a denormalized doc store for their technical needs, and the other because it has an easier quickstart on day 1 when you don’t know SQL. Here we are usually talking of the latter. The beauty of something like Postgres is that you don’t need to get your schema right on day 1. How could you anyways unless tou already know all future data access pattern?

kodablah 8y ago

For me, it's two main reasons: 1) Generating and storing massive amounts of pre-computed permutations of the same data based on how I'll read it with easy linear growth by adding servers and 2) Transparent multi-master reads/writes providing easy HA failover. Note, not all NoSQL options offer this. If existing free relational DBs had these options easy and out of the box, I would use them instead. I just need a ton of data, often redundant, that I never delete, that supports distributed writes and reads (ACID not required). The things I use Cassandra for would require huge Postgres servers and custom sharding/partitioning setups and Citus HA and all sorts of other things.

njharman 8y ago

> what it's like not to have the ability to manipulate the database. In general "not wanting to understand, depend on, or work with other systems/teams" (often cause of dysfunctional eng/pm organizations) has, over 20years experience, been the number one reason devs (esp front end devs cause they tend to move the fastest (faster than traditional backend teams are prepared for) and have least experience in back end systems) go for something new, something they can control, something they can spin up fast and quick. NoSQL, cloud, anything they can do themselves.