Problem with assert referenced columns are present in datasource

[lilila]

@villebro , I have an issue with #10451

It breaks the following type of jinja-based filters:

{% set min_tracking  = filter_values('min_sem_tracking', -9 )[0] %} 
{% set max_tracking = filter_values('max_sem_tracking', 2 )[0]  %} 


SELECT 
...
WHERE 
sem_tracking > {min_tracking} AND sem_tracking < {max_tracking} 

[issue-label-bot]

Issue-Label Bot is automatically applying the label #bug to this issue, with a confidence of 0.96. Please mark this comment with 👍 or 👎 to give our bot feedback!

Links: app homepage, dashboard and code for this bot.

[villebro]

Thanks for posting this issue @lilila . Can you post some more details on how to reproduce the error?

[lilila]

Hi Villebro,
If I want to add a min / max filter on a field (since the filter box does not allow this) I usually create a SQL Lab view as follows :

SELECT -5 as min_sem_tracking, 1 as max_sem_tracking
Then I create a filter box with the 2 fields created in this "fake" table :

After adding the filter box in the dashboard I use the user input value for min and max value to create more "complex" filters on another table. As explain above I collect user inputs using filter_values

{% set min_tracking  = filter_values('min_sem_tracking', -9 )[0] %} 
{% set max_tracking = filter_values('max_sem_tracking', 2 )[0]  %} 

Then I use the collected values to create filters on another field in another table :

SELECT * FROM my_table 
WHERE 
sem_tracking > {min_tracking} AND sem_tracking < {max_tracking} 

With the new restriction, I receive an error message saying min_sem_tracking and max_sem_tracking are not in the datasource.

[villebro]

I was able to reproduce, thanks @lilila and @bkyryliuk for helping track this down. Working on a fix now.

[lilila]

Hi @villebro , I hope you are doing well.
I tried the new release (1.0.rc1) and I have the impression the problem is back. All "artificial filters " (such as min / max values) generate warnings. In the end I end up with a dashboard looking like this :

[villebro]

@lilila is the behavior consistent? There is a problem that causes chart loading errors on 1.0.0rc1 that should be fixed by this PR: #12489