What is the correct/ smart way to build two separate datasets (tables) where are both common and not common columns so that common columns can be edited in either of the datasets and they stay in sync automatically? And also: the common columns carry one entity per row, like: “ID” + “name” + “date” + “status”, which should stay coherent regardless of edits from both ‘directions’
This is simple for a single linked column use case but is it possible at all for multiple dependent columns?
Also, I know that one (and perhaps the correct) answer is that all data should be in a single massive table and then just publish parallel subsets of that single table to separate views (pages). This would work but reason to ask is the need to try to make it “easier for humans to maintain”.
Best,
Tomas
Hi Tom, as far as I understand, you must have a “single source of truth”. That means these columns that you want to appear in either data set should NOT be on both datasets. Only one dataset (table) should have that column. The other dataset must only reference/show that column from the other data set.
Since you didnt provide a user case, let me create a fictitious one.
You have clients and their data. That’s a table. Then you have projects and you want to see several data about a client on the Project table. But let’s say part of the project is to gather info about the client, so you in the project table you may have to change some info about the client.
The client column is a reference to the client table. You can´t CHANGE it in the Projects Table.
HOWEVER if the user has permission, the user can click at the little CLIP that indicates a refence and it will open that record AS IF IT WAS A POP-UP window. And that window showing the Client Record can be edited.
@Rogerio_Penna Ah! My bad, I did not read the manual. Just thought the clip is to illustrate the fact it is a reference 
Still learning…
And ‘one source of truth’ is a must principle of course.
So, I could reference multiple columns from one table to another and they all would be editable as you described also from the referencing direction? Do they also let user to create new input from referencing direction? (I can of course test that myself).
And to conclude - I already kind of decided to go for a large single table as it makes sense for this use case and won’t bring the issue of parallel items. The way to make it more usable will be the widgets. I need to make many pages and then use some HTML to give nice “top tab bar menu” to pick individual pages which each have a subset of the large table + they have a widget with card view for easier viewing and reading of large table column specific things.
Thanks for the input. Let’s mark this completed.
1 Like