[AURON #1502] Implement native function of bround.

[slfan1989]

Which issue does this PR close?

Closes #1502.

Rationale for this change

To achieve full compatibility with Spark’s numeric functions, we should implement bround() with the following characteristics:

Expected behavior

Function name: bround(expr, scale)
Rounding mode: HALF_EVEN (bankers’ rounding)
Example:

• bround(2.5) → 2.0
• bround(3.5) → 4.0
• bround(-2.5) → -2.0

Supports: FLOAT, DOUBLE, DECIMAL, INT16/32/64

• Handles negative scales: e.g., bround(123.4, -1) = 120
• Null-safe: should return NULL if input is NULL
• Array and scalar inputs: consistent with current round() implementation

What changes are included in this PR?

This PR adds full support for the bround() function, which performs bankers’ rounding (HALF_EVEN).
The following changes are included:

• Added native implementation of spark_bround() in the expression layer.
• Added BRound expression support in NativeConverters for proper Spark → native translation.
• Added comprehensive unit tests to verify correctness for:
  • Positive and negative numbers
  • Different scale values (including negative scales)
  • Various numeric types (FLOAT, DOUBLE, DECIMAL, INT, LONG)
  • Null input handling

Are there any user-facing changes?

No.

How was this patch tested?

CI.

[slfan1989]

@richox Could you please help review this PR? Thank you very much!

[slfan1989]

@richox Could you please help merge this PR into the main branch? Thank you very much! PR #1502 has become stale.

[slfan1989]

@richox Thank you very much for the review and merge!