Add more tests showing coercing behavior with literals

[alamb]

Which issue does this PR close?

• Follow on to fix(expr-common): Coerce to Decimal(20, 0) when combining UInt64 with signed integers #14223
• Follow on to Minor: Add tests for types of arithmetic operator output types #14250

Rationale for this change

I was still confused about what happens with coercion in #14223 (comment) and somewhat panic'ing that the change would impact performance. I want tests in place to show the output types of operations

What changes are included in this PR?

Add more sqllogictests showing the output types when adding literals to various column types

Are these changes tested?

Are there any user-facing changes?

[alamb]

This shows that adding a literal integer (2) does not upcast to Decimal for integer types (which would be terrible for performance)

[ozankabak]

I think adding tests for comparison operations will probably expose the possible issue with the linked PR. (Or give us a peace of mind)

[alamb]

This test shows that with a positive integer comparison does not result in any casting of the column

[alamb]

This test show that with a negative integer, the comparison to int64 does not cast.

However, a comparison with -5 does cast to Decimal128: CAST(uint64@7 AS Decimal128(20, 0)) < Some(-5),20,0

When I reverted the changes in #14223 locally, this is the difference (cast to Int64 rather than Decimal):

[Diff] (-expected|+actual)
    physical_plan
    01)CoalesceBatchesExec: target_batch_size=8192
-   02)--FilterExec: int64@3 < -5 AND CAST(uint64@7 AS Decimal128(20, 0)) < Some(-5),20,0 AND float64@9 < -5 AND decimal@10 < Some(-500),5,2
+   02)--FilterExec: int64@3 < -5 AND CAST(uint64@7 AS Int64) < -5 AND float64@9 < -5 AND decimal@10 < Some(-500),5,2
    03)----MemoryExec: partitions=1, partition_sizes=[1]
at test_files/operator.slt:286

However, note that in this case the predicate will never be true anyways (a uint64 can never be less than -5 as it is by definition always greater than zero -- something @gatesn pointed out on another ticket)

[alamb]

I think adding tests for comparison operations will probably expose the possible issue with the linked PR. (Or give us a peace of mind)

It is a good idea -- I added tests showing what happens in comparisons

[jonahgao]

Thanks @alamb . I think these tests are very valuable; they reveal the problems more clearly.

[alamb]

As this is entirely tests and demonstrates the actual behavior I am going to merge it as is -- I am happy to add more tests if we discover additional ones to add

[alamb]

Thanks again @jonahgao