fix: Avoid integer overflow in split_part() (#20198)

Along the way, improve the docs slightly.

## Rationale for this change

Evaluating `SELECT SPLIT_PART('', '', -9223372036854775808);` yields (in
a debug build):

```
thread 'main' (41405991) panicked at datafusion/functions/src/string/split_part.rs:236:47:
attempt to negate with overflow
```

<!--
Why are you proposing this change? If this is already explained clearly
in the issue then this section is not needed.
Explaining clearly why changes are proposed helps reviewers understand
your changes and offer better suggestions for fixes.
-->

## Are these changes tested?

Yes, added unit test.

<!--
We typically require tests for all PRs in order to:
1. Prevent the code from being accidentally broken by subsequent changes
2. Serve as another way to document the expected behavior of the code

If tests are not included in your PR, please explain why (for example,
are they covered by existing tests)?
-->

## Are there any user-facing changes?

<!--
If there are user-facing changes then we may require documentation to be
updated before approving the PR.
-->

<!--
If there are any breaking changes to public APIs, please add the `api
change` label.
-->

Author: neilconway

datafusion/functions/src/string/split_part.rs

@@ -48,7 +48,10 @@ use std::sync::Arc;
 ```"#,
     standard_argument(name = "str", prefix = "String"),
     argument(name = "delimiter", description = "String or character to split on."),
-    argument(name = "pos", description = "Position of the part to return.")
+    argument(
+        name = "pos",
+        description = "Position of the part to return (counting from 1). Negative values count backward from the end of the string."
+    )
 )]
 #[derive(Debug, PartialEq, Eq, Hash)]
 pub struct SplitPartFunc {
@@ -233,7 +236,7 @@ where
                         std::cmp::Ordering::Less => {
                             // Negative index: use rsplit().nth() to efficiently get from the end
                             // rsplit iterates in reverse, so -1 means first from rsplit (index 0)
-                            let idx: usize = (-n - 1).try_into().map_err(|_| {
+                            let idx: usize = (n.unsigned_abs() - 1).try_into().map_err(|_| {
                                 exec_datafusion_err!(
                                     "split_part index {n} exceeds minimum supported value"
                                 )
@@ -324,6 +327,20 @@ mod tests {
             Utf8,
             StringArray
         );
+        test_function!(
+            SplitPartFunc::new(),
+            vec![
+                ColumnarValue::Scalar(ScalarValue::Utf8(Some(String::from(
+                    "abc~@~def~@~ghi"
+                )))),
+                ColumnarValue::Scalar(ScalarValue::Utf8(Some(String::from("~@~")))),
+                ColumnarValue::Scalar(ScalarValue::Int64(Some(i64::MIN))),
+            ],
+            Ok(Some("")),
+            &str,
+            Utf8,
+            StringArray
+        );
 
         Ok(())
     }

docs/source/user-guide/sql/scalar_functions.md

@@ -1891,7 +1891,7 @@ split_part(str, delimiter, pos)
 
 - **str**: String expression to operate on. Can be a constant, column, or function, and any combination of operators.
 - **delimiter**: String or character to split on.
-- **pos**: Position of the part to return.
+- **pos**: Position of the part to return (counting from 1). Negative values count backward from the end of the string.
 
 #### Example