Build application in REI3, on a previously existing posgreSQL schema.

rfernandez

Hi,
How are you?

I have a postgres schema already loaded with data. This schema was created directly in postgres.
How can I build a REI3 application on that schema?
From what I could see, besides creating the postgres schema and tables for the application, REI stores some metadata and definitions about the database objects in its instance tables (apologies if I'm not using the correct names, but I hope the scenario and the question are clear enough).

Thank you !!
Warm regards

gabriel

Short answer: You cannot.

The r3 system must be in control of the database and the corresponding DB schemas, as there are many references and updates to be done for any change in the DB. While it is technically possible to do that outside of r3, it would be complicated and error prone.

What is the reason you need to run a r3 app from inside a pre-existing DB schema?

rfernandez

Ok. I understand.

Let me clarify just in case...
What I want is to avoid recreating from scratch, a whole postgres schema I already have (both structure and data) from within REI3, which as of my current knowledge of the platform would imply defining each table/relation, the columns, the relationships between the relations, and then creating each record or somehow bulk loading the tables from a csv file, or something similar. (perhaps there is a way and I just don't currently know about it)

It would be very practical to be able to do such thing as "ingesting a schema" into a REI3 format (adding all the required metadata), as there are many schemas already existing (some with tens or even hundreds of relations, fields, etc., and thousands of records) that just need an application on top to work with them, or do some charting or reporting. (for which REI3 excels)

I hope now the scenario is a bit more clear.

And as if you said, it would not be recommended to do this manually, what would you suggest to minimize the effort to get to the point in which you have all this structures/data ready to work with r3?

Thank you!!
Best regards

gabriel

Thank you for going into detail about your use case. Unfortunately, this is far from simple.

While it is technically possible, there are many permutations of possible configuration in Postgres for tables & columns that we neither can nor want to support. And if we cannot support something, it would either not work as expected or just be omitted.

Doing this would also enable people to use r3 as frontend for other tools working in the same Postgres DB. This is quite error-prone and, if not handled very carefully, would be a recipe for disastrous data loss.

r3, like many low code tools, is an integrated platform - frontend & backend. Unless you know exactly what you do - and I mean deeply know what all the components do, including all levels of caches involved - I would highly discourage this kind of solution. Not just related to r3, but to all low code tools in general. Unless designed to be a loosely coupled frontend to external DBs, low code tools assume they have control over their backend.

We aim to design solutions for stable business operations. This kind of thing, unless specifically designed for, is very finicky and not something we are interested in offering.

rfernandez

gabriel
Ok. I undestand your point.

Thank you!