[AURON #1724] Support binary input for Spark substring function

[lyne7-sc]

Which issue does this PR close?

Closes #1724

Rationale for this change

Spark substring supports both string and binary inputs, while Auron previously mapped it to datafusion's Substr, which only handled string-compatible behavior and caused the Spark string/binary substring suite case to be excluded.

What changes are included in this PR?

• Add native Spark_Substring ext function support for Utf8 and Binary inputs.
• Route Spark Substring conversion through Spark_Substring and preserve the input data type.
• Add native unit coverage for string and binary substring.
• Re-enable string / binary substring function test case.

Are there any user-facing changes?

Yes. Spark SQL substring now supports binary input in native execution, matching Spark behavior.

How was this patch tested?

• Added native unit tests for spark_substring.
• Spark String Function Suite

[ShreyeshArangath]

Mostly LGTM, left a few comments

[lyne7-sc]

@ShreyeshArangath Thanks for your review! I’ve updated the implementation accordingly.

[ShreyeshArangath]

Great work! LGTM, just one question

[yew1eb]

The Rust implementation looks correct. One issue on the Scala side.

[yew1eb]

The guard pos > 0 && len >= 0 was inherited from the old Substr mapping and is now too restrictive. spark_substring correctly handles pos == 0 (treated as 1) and negative pos (counted from the end), as the Rust unit tests already cover. Cases with pos <= 0 will still fall back to Spark instead of using the native path. Consider relaxing the guard to just len >= 0, or dropping it entirely since the implementation handles all integer values.

[lyne7-sc]

Good catch! That makes sense to me. I’ve relaxed the guard so these cases can go through the native path.

[ShreyeshArangath]

Great work!