The question I have is this, and don’t attack me for asking, but do people still produce database diagrams for use at work? I thought we had abstracted this by using Domain Driven design, ORMs, APIs, and the like.
Do individual teams still document their tables or do they document their entities? Do you still hand roll sql in a dao or do you use some higher abstractions?
Curious minds want to know because it’s been over a decade since I’ve done any database design documentation and have only done lean relationships and domain modeling documentation. Swaggering the rest.
I will say this, I love the look of this and I would love to just draw abstract shapes and things like I do on Miro. AWS architectures, etc etc.
We do at my team.
I do not work on the Backend side, but I help in designing the DB schema.
Whenever I am thinking of new features to be developed, or when engineers are suggesting some features/approaches, looking at the ERD helps a ton. Onboarding is also easier.
We were using Lucidchart[1] until we reached the limit[2], so we found dbdiagram.io (which is just not the same).
[1] So far, it was the best of the market in terms of "canvas freedom"
[2] We are struggling with salaries, so we are saving everywhere
> Also, Lucidchart is just draw.io fyi - you don't need a license unless you need the file sharing aspect, and I feel like it belongs closer to the code.
No, it is not. I took the whole day to check what is available in the market, and concluded that Lucidchart UX is quite balanced.
well, they used to be - I just went and saw they sprinkled llms into it so yeah, now it's different. Diagrams.net aka Draw.io was what they were using...
We just assume 6th order and swing for the fences? Why do you think that is? Because we have too much data? I remember a guy who was in charge of schema but that was as close as we got to documenting the actual database.
I'd love to hear more about your experience with AGPL licensing and building a cloud/SaaS product later. It looks like you introduced a CLA four months ago, but have a lot of external contributions prior to that.
How did you solve the AGPL hurdles for that pre-CLA third-party code? It's usually impractical to get a lot of contributors to retroactively sign a CLA. Did you have to rewrite some/all of those older code contributions, or is the full cloud product open source as well?
Been using ChartDB recently and it’s been quite good for my use case. I’m on Supabase, and found that the Supabase built-in ERD viewer kind of breaks down once you go past like 15 tables.
I tried a few other DB visualization tools but was turned off because it felt like I needed to learn another toolset, when I just wanted to import my schema and start working with it. ChartDB's approach was better for me, the single query import worked fine, and adding to the schema works how you'd expect.
Really hoping to see a feature for “diffs between the dev diagram and the live DB”. we’re collaborating on the diagram as a team, and it’d be great to track what changed between our current draft and the actual database.
Pretty.
The question I have is this, and don’t attack me for asking, but do people still produce database diagrams for use at work? I thought we had abstracted this by using Domain Driven design, ORMs, APIs, and the like.
Do individual teams still document their tables or do they document their entities? Do you still hand roll sql in a dao or do you use some higher abstractions?
Curious minds want to know because it’s been over a decade since I’ve done any database design documentation and have only done lean relationships and domain modeling documentation. Swaggering the rest.
I will say this, I love the look of this and I would love to just draw abstract shapes and things like I do on Miro. AWS architectures, etc etc.
We do at my team. I do not work on the Backend side, but I help in designing the DB schema.
Whenever I am thinking of new features to be developed, or when engineers are suggesting some features/approaches, looking at the ERD helps a ton. Onboarding is also easier.
We were using Lucidchart[1] until we reached the limit[2], so we found dbdiagram.io (which is just not the same).
[1] So far, it was the best of the market in terms of "canvas freedom" [2] We are struggling with salaries, so we are saving everywhere
how many tables? is it manageable and interesting like trying to squeeze 2b records into postgres or what?
Also lucidchart is just draw.io fyi - you don't need a license unless you need the file sharing aspect and I feel like it belongs closer to the code.
I'm just generally curious what others' process is for this as the only meetings I've had in the last 5 years were who was getting cut.
If I recall, Lucidchart allows up to 15 "shapes.
> Also, Lucidchart is just draw.io fyi - you don't need a license unless you need the file sharing aspect, and I feel like it belongs closer to the code.
No, it is not. I took the whole day to check what is available in the market, and concluded that Lucidchart UX is quite balanced.
well, they used to be - I just went and saw they sprinkled llms into it so yeah, now it's different. Diagrams.net aka Draw.io was what they were using...
The databases often out live the abstraction.
Yes, very much so but I think it's probably a self fulfilling thing?
If I care about the data model so much that I want to diagram it then it's going to be the exact DB schema.
People have stopped documenting their databases, which is a loss for the industry.
We just assume 6th order and swing for the fences? Why do you think that is? Because we have too much data? I remember a guy who was in charge of schema but that was as close as we got to documenting the actual database.
only for learning
for industry???? let me create an ERD for my 10th SaaS tools that need generic auth and payment, nope
I'd love to hear more about your experience with AGPL licensing and building a cloud/SaaS product later. It looks like you introduced a CLA four months ago, but have a lot of external contributions prior to that.
How did you solve the AGPL hurdles for that pre-CLA third-party code? It's usually impractical to get a lot of contributors to retroactively sign a CLA. Did you have to rewrite some/all of those older code contributions, or is the full cloud product open source as well?
Been using ChartDB recently and it’s been quite good for my use case. I’m on Supabase, and found that the Supabase built-in ERD viewer kind of breaks down once you go past like 15 tables.
I tried a few other DB visualization tools but was turned off because it felt like I needed to learn another toolset, when I just wanted to import my schema and start working with it. ChartDB's approach was better for me, the single query import worked fine, and adding to the schema works how you'd expect.
Really hoping to see a feature for “diffs between the dev diagram and the live DB”. we’re collaborating on the diagram as a team, and it’d be great to track what changed between our current draft and the actual database.
Any thoughts of supporting .dot files?