Unexpected data type when conditionally updating column

[frankvollebregt]

I've written a function in my .drift file to update a certain table. One of the columns that is updated is the status, which should only ever become larger (and can be ignored if a lower status is pushed). I still want to update the other column values though, so UPDATE ... WHERE would not work in my case. I came across a StackOverflow post which tries to solve a similar issue. However, when I implement it like this

updateStatus: UPDATE my_table SET status = MAX(status, :newStatus), second_col = :secondCol;

and generate my function, it thinks newStatus should be a String, while status is defined like this:

CREATE TABLE my_table (
  ...
  status INT NULL
);

IDE error message: The argument type 'int' can't be assigned to the parameter type 'String'.

And indeed when I look in the generated code, it has set the type to String:

variables: [
  Variable<String>(newStatus),
  ...
]

[North101]

You need to give :newStatus a type as it seems its not inferred.

updateStatus(:newStatus AS INT):
UPDATE my_table SET status = MAX(status, :newStatus), second_col = :secondCol;

https://drift.simonbinder.eu/docs/using-sql/drift_files/#variables

[frankvollebregt]

Ah, I did not know I could mark them as a certain type. Thank you!

[simolus3]

I've also fixed the inaccurate analysis in 645489a. If you notice more cases like this, please let me know :)

[davidmartos96]

@simolus3 Related to this, I found the following warning a while back.
When using the expression strftime('%s','now') * 1000, it warns about strftime not being numeric. %s returns the number of seconds since epoch. https://www.sqlite.org/lang_datefunc.html