I think I was able to isolate my problem with Query column but now I don’t know if this behavior is normal by design or a real bug.
This my data to test, the data is generated by API request fired every 30 sec and of course, the data is going to change. Due to values depend on their timestamp, their positions (row) are different on each API request (30 sec) from top to bottom…
In this case, when the relation is first created, the results are correct but when the data is changed by a new API request (every 30 sec), the relation freezes and the results are not updated.
My question is: Can a Query column work with a computed column as filter? I hope so!
First pass thru the test app and is rendering correct AND my app (using lots and lots of compute columns and lookup tables) is rendering correctly.
I have about 25 more dashboard data instances I can check later this afternoon which are all based on computed columns and lookups and will verify they render correctly for my use case - SO FAR SO GOOD.
BTW - the UI for query is off-the-charts excellent - both intuitive and functional. Congrats to engineering and product management and the UI team for the great work.
I tested all, here are the results. I had built ‘tags’ to match against so I could use relations to ‘roll-up’ to create the ‘scorecards’. Hundreds and hundreds of tags so I could match WEEK, MONTH, QUARTER, YEAR, CHANNEL, PARTNER, SOURCE, TERRITORY and their permutations (using array matching which is also fantastic). All to get to the 60 or so data points I wanted to show plus ‘future’ reporting.
The TAG method was very labor intensive and error prone which I just discovered.
I moved all the ‘scorecards’ to Queries and just checked the two results. I found 2 errors in the TAG method (wrong template data and wrong logic) and 1 error in the Query method (I need another filter to get the correct result).
Using Queries was a 50x time/logic improvement over the TAG method. My only concern is performance - the TAG method was fast (but also ridiculously unmanageable). The query method on my small data sets is fine and I am sure that Glide will continue to improve its performance. It is a game-changing feature and so intuitive.
@david and @Mark - fantastic execution by the entire Glide team.
I will wait for others to chime in but from my perspective I mark this as the solution.
