5 ms·

I recall when my team chose MongoDB ~2011, Postgres & friends didn't have JSON columns, so there was a lot of extra data modelling that probably was unnecessary

I recall when my team chose MongoDB ~2011, Postgres & friends didn't have JSON columns, so there was a lot of extra data modelling that probably was unnecessary. The biggest use case for MongoDB was for huMongous data. Obvs MongoDB was a good fit, because of the name.

jerf 3y ago
Yes, I think it's reasonable to say the NoSQL community drove the JSON columns in the relational databases. And while the JSON-ness of said columns is not necessarily the relevant bit to me, having the escape hatch of being able to just sort of splat out a "other" column has been something I've used several times, and it has relieved some non-trivial stress and strain in relational designs in my opinion. It's not even that hard to migrate something out of the JSON column if you want to raise it to a full column later, and JSON indices mean you may not even really need to. I have a hard time justifying a document database in 2024 when I have 2024 Postgres; I don't know how large the window is where MongoDB will do something Postgres won't, but I know my systems are not in that window. (Assuming it even exists, which I wouldn't be surprised it doesn't, but for the sake of argument I will assume it does.) But a decade ago that would be a much harder decision if I truly had a document-based use case.
- jimbokun 3y ago
  Horizontal scaling is still challenging in Postgres. That’s where “NoSQL” systems still have their place. Or one of the SQL systems layered on to a NoSQL architecture, with the trade offs that entails.
wlll 3y ago
https://www.postgresql.org/about/featurematrix/#json https://www.postgresql.org/about/featurematrix/#json You're right, 9.4/9.5 were when JSONB was introduced and expanded (release 2013/2014). That said, it's a big decision to go with two very very different technologies (relational vs Document store) just for a column type, storing JSON must have been a pretty major product feature?
vbezhenar 3y ago
I still didn't have a single use-case to use JSON columns yet. Sometimes I feel that JSON columns are made for people who don't want to properly design their domain model. Big mistake.
- abound 3y ago
  The main use case I find myself reaching for JSON/JSONB fields is unstructured map-like metadata, like KV tags on a resource, where the data model/domain has no expectation that any given field is set.
- unclebucknasty 3y ago
  Interesting point, but I wonder if there's a little confirmation bias and/or circular logic there. In other words that you don't see a use case for JSON columns because you don't think it's good DB design. I know I have used them after decades of RDBMS work, and it feels like I'm "cheating" a bit. But they really do make sense for some use cases that can be harder to achieve with traditionally normalized RDBMS alone. For instance, for configuration options or scenarios where different objects may have a few different attributes, but are largely the same. If you've ever used an entity-attribute-value data model or otherwise ended up with table proliferation or a bunch of nullable columns to accommodate the latter use case, you might appreciate the simplification JSON columns offer.
- onetimeuse92304 3y ago
  I think it is the same misguided discussion as strongly, staticall typed vs weakly typed languages. (Mind I mean the discussion is misguided, not you) The problem is that relational model is a one solution to certain problem of when you want to pay the price of enforcing the model of your data. Relational database is when you pay all the cost upfront. With MongoDB I have other options. It does not mean they are necessarily better options. But some of them are better in some situations, for example when rapid prototyping. Or when you need to preserve data in different formats as the application evolves. Personally, if you can, you probably should model your data upfront. But it is good to have other options, sometimes.

jerf 3y ago

Yes, I think it's reasonable to say the NoSQL community drove the JSON columns in the relational databases. And while the JSON-ness of said columns is not necessarily the relevant bit to me, having the escape hatch of being able to just sort of splat out a "other" column has been something I've used several times, and it has relieved some non-trivial stress and strain in relational designs in my opinion. It's not even that hard to migrate something out of the JSON column if you want to raise it to a full column later, and JSON indices mean you may not even really need to. I have a hard time justifying a document database in 2024 when I have 2024 Postgres; I don't know how large the window is where MongoDB will do something Postgres won't, but I know my systems are not in that window. (Assuming it even exists, which I wouldn't be surprised it doesn't, but for the sake of argument I will assume it does.) But a decade ago that would be a much harder decision if I truly had a document-based use case.

jimbokun 3y ago
Horizontal scaling is still challenging in Postgres. That’s where “NoSQL” systems still have their place. Or one of the SQL systems layered on to a NoSQL architecture, with the trade offs that entails.

wlll 3y ago

https://www.postgresql.org/about/featurematrix/#json https://www.postgresql.org/about/featurematrix/#json You're right, 9.4/9.5 were when JSONB was introduced and expanded (release 2013/2014). That said, it's a big decision to go with two very very different technologies (relational vs Document store) just for a column type, storing JSON must have been a pretty major product feature?

vbezhenar 3y ago

I still didn't have a single use-case to use JSON columns yet. Sometimes I feel that JSON columns are made for people who don't want to properly design their domain model. Big mistake.

abound 3y ago
The main use case I find myself reaching for JSON/JSONB fields is unstructured map-like metadata, like KV tags on a resource, where the data model/domain has no expectation that any given field is set.
unclebucknasty 3y ago
Interesting point, but I wonder if there's a little confirmation bias and/or circular logic there. In other words that you don't see a use case for JSON columns because you don't think it's good DB design. I know I have used them after decades of RDBMS work, and it feels like I'm "cheating" a bit. But they really do make sense for some use cases that can be harder to achieve with traditionally normalized RDBMS alone. For instance, for configuration options or scenarios where different objects may have a few different attributes, but are largely the same. If you've ever used an entity-attribute-value data model or otherwise ended up with table proliferation or a bunch of nullable columns to accommodate the latter use case, you might appreciate the simplification JSON columns offer.
onetimeuse92304 3y ago
I think it is the same misguided discussion as strongly, staticall typed vs weakly typed languages. (Mind I mean the discussion is misguided, not you) The problem is that relational model is a one solution to certain problem of when you want to pay the price of enforcing the model of your data. Relational database is when you pay all the cost upfront. With MongoDB I have other options. It does not mean they are necessarily better options. But some of them are better in some situations, for example when rapid prototyping. Or when you need to preserve data in different formats as the application evolves. Personally, if you can, you probably should model your data upfront. But it is good to have other options, sometimes.