Consolidate Example: simplify_udaf_expression.rs into advanced_udaf.rs

[alamb]

Is your feature request related to a problem or challenge?

• Part of [EPIC] Improve examples to make them easier to navigate #11172

The large number of small examples in DataFusion leads to multiple problems:

• Unnecessarily large amounts of temporary disk space (each example binary takes 100+ MB)
• Larger cargo t time (each binary must be linked / created)
• Makes it harder to find the relevant examples

Thus I think the fewer examples we have the

1. faster our dev experience will be
2. easier it will be to find the relevant examples

Describe the solution you'd like

• Consolidate the contents of simplify_udaf_expression
• Into advanced_udaf.rs

Then update the readme page:

• https://github.com/apache/datafusion/blob/main/datafusion-examples/README.md

Describe alternatives you've considered

No response

Additional context

I think this is a good first issue as the task is straightforward and doesn't require deep understanding of the DataFusion codebase

[takaebato]

take

[alamb]

Thank you @takaebato 🙏

[takaebato]

@alamb

Sorry for the delay🙏 But I have a quick question!
I’m not too sure about this, but should the return type of AggregateFunctionSimplification perhaps be Result<ExprSimplifyResult>(or a similar enum that can indicate whether or not simplification occurred)?

Currently, it doesn’t seem possible to skip simplification, as it always returns a simplified Expr (or Err).
The simplify method can return an Option, but the decision to simplify seems to rely on information passed to AggregateFunctionSimplification, so might it be better if this method can choose whether or not to simplify..?

[alamb]

@alamb

Sorry for the delay🙏 But I have a quick question! I’m not too sure about this, but should the return type of AggregateFunctionSimplification perhaps be Result<ExprSimplifyResult>(or a similar enum that can indicate whether or not simplification occurred)?

Currently, it doesn’t seem possible to skip simplification, as it always returns a simplified Expr (or Err). The simplify method can return an Option, but the decision to simplify seems to rely on information passed to AggregateFunctionSimplification, so might it be better if this method can choose whether or not to simplify..?

The rationale for returning ExprSimplifyResult is so that the argument doesn't have to be cloned -- aka to avoid coping the ownership of argument is passed to the function and there needs to be some way to get the original structuer back

[takaebato]

@alamb

Thank you so much for the explanation!!
(I also realized that if we want to use the actual implementation conditionally, we can return Ok(Expr::AggregateFunction(aggregate_function)).)