Fix string array equality when the values buffer is the same and only the offsets to access it differ (apache#9325)

# Which issue does this PR close?

<!--
We generally require a GitHub issue to be filed for all bug fixes and
enhancements and this helps us generate change logs for our releases.
You can link an issue to this PR using the GitHub syntax.
-->

- Closes apache#9323.

# Rationale for this change

This bug seems to have existed for at least the last 5 years (before
that, array equality had a lot more issues with offsets).

<!--
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.
-->

# What changes are included in this PR?

<!--
There is no need to duplicate the description in the issue here but it
is sometimes worth providing a summary of the individual changes in this
PR.
-->

# Are these changes tested?

Tested with unit tests.

# Are there any user-facing changes?

no

Author: jhorstmann

arrow-array/src/array/string_array.rs

@@ -551,4 +551,12 @@ mod tests {
         let err_return = array.into_builder().unwrap_err();
         assert_eq!(&err_return, &shared_array);
     }
+
+    #[test]
+    fn test_non_null_string_array_equal() {
+        let a = StringArray::from(vec![Some("ab"), Some("c")]);
+        let b = StringArray::from(vec![Some("a"), Some("bc")]);
+
+        assert_ne!(a, b);
+    }
 }

arrow-data/src/equal/list.rs

@@ -21,7 +21,7 @@ use num_integer::Integer;
 
 use super::equal_range;
 
-fn lengths_equal<T: ArrowNativeType + Integer>(lhs: &[T], rhs: &[T]) -> bool {
+pub(super) fn lengths_equal<T: ArrowNativeType + Integer>(lhs: &[T], rhs: &[T]) -> bool {
     // invariant from `base_equal`
     debug_assert_eq!(lhs.len(), rhs.len());
 

arrow-data/src/equal/variable_size.rs

@@ -15,12 +15,12 @@
 // specific language governing permissions and limitations
 // under the License.
 
+use super::utils::equal_len;
 use crate::data::{ArrayData, contains_nulls};
+use crate::equal::list::lengths_equal;
 use arrow_buffer::ArrowNativeType;
 use num_integer::Integer;
 
-use super::utils::equal_len;
-
 fn offset_value_equal<T: ArrowNativeType + Integer>(
     lhs_values: &[u8],
     rhs_values: &[u8],
@@ -63,15 +63,18 @@ pub(super) fn variable_sized_equal<T: ArrowNativeType + Integer>(
     // Only checking one null mask here because by the time the control flow reaches
     // this point, the equality of the two masks would have already been verified.
     if !contains_nulls(lhs.nulls(), lhs_start, len) {
-        offset_value_equal(
-            lhs_values,
-            rhs_values,
-            lhs_offsets,
-            rhs_offsets,
-            lhs_start,
-            rhs_start,
-            len,
-        )
+        let lhs_offsets_slice = &lhs_offsets[lhs_start..lhs_start + len + 1];
+        let rhs_offsets_slice = &rhs_offsets[rhs_start..rhs_start + len + 1];
+        lengths_equal(lhs_offsets_slice, rhs_offsets_slice)
+            && offset_value_equal(
+                lhs_values,
+                rhs_values,
+                lhs_offsets,
+                rhs_offsets,
+                lhs_start,
+                rhs_start,
+                len,
+            )
     } else {
         (0..len).all(|i| {
             let lhs_pos = lhs_start + i;
@@ -95,3 +98,35 @@ pub(super) fn variable_sized_equal<T: ArrowNativeType + Integer>(
         })
     }
 }
+
+#[cfg(test)]
+mod tests {
+    use crate::ArrayData;
+    use crate::equal::variable_size::variable_sized_equal;
+    use arrow_buffer::Buffer;
+    use arrow_schema::DataType;
+
+    #[test]
+    fn test_variable_sized_equal_diff_offsets() {
+        let a = ArrayData::builder(DataType::Utf8)
+            .buffers(vec![
+                Buffer::from_vec(vec![0_i32, 3, 6]),
+                Buffer::from_slice_ref(b"foobar"),
+            ])
+            .null_bit_buffer(Some(Buffer::from_slice_ref([0b01_u8])))
+            .len(2)
+            .build()
+            .unwrap();
+        let b = ArrayData::builder(DataType::Utf8)
+            .buffers(vec![
+                Buffer::from_vec(vec![0_i32, 2, 6]),
+                Buffer::from_slice_ref(b"foobar"),
+            ])
+            .null_bit_buffer(Some(Buffer::from_slice_ref([0b01_u8])))
+            .len(2)
+            .build()
+            .unwrap();
+
+        assert!(!variable_sized_equal::<i32>(&a, &b, 0, 0, 2));
+    }
+}