You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is your feature request related to a problem or challenge?
For a brief time we considered it necessary to extend StateFieldArgs to include the nullable derived from the schema of the first argument. So that we could define a list item for the accumulator state like this:
// The data type and the `nullable` is equivalent to schema of the first argumentField::new("item", args.input_type.clone(), args.input_nullable)
Note: The field input_nullable in the above code does not exist in StateFieldArgs.
Instead we currently do this in code:
Field::new("item", args.input_type.clone(),true)// allows null values in aggregate computation
We take advantage of the fact that currently aggregates do no have any special treatment for null values present inside list of accumulator state. So we can avoid adding a new field like input_nullable to StateFieldArgs.
Describe the solution you'd like
Capture this decision explicitly as a doc comment to benefit future readers of code. Add the comment to the following aggregates:
Is your feature request related to a problem or challenge?
For a brief time we considered it necessary to extend
StateFieldArgs
to include thenullable
derived from the schema of the first argument. So that we could define a list item for the accumulator state like this:Note: The field
input_nullable
in the above code does not exist inStateFieldArgs
.Instead we currently do this in code:
We take advantage of the fact that currently aggregates do no have any special treatment for null values present inside list of accumulator state. So we can avoid adding a new field like
input_nullable
toStateFieldArgs
.Describe the solution you'd like
Capture this decision explicitly as a doc comment to benefit future readers of code. Add the comment to the following aggregates:
sum
count
array_agg
nth_value
bit_and_or_xor
There are no code changes required.
Describe alternatives you've considered
No response
Additional context
See #11433 (comment)
The text was updated successfully, but these errors were encountered: