Refactor ListArray hashing to consider only sliced values (#19500)

## 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. For example
`Closes #123` indicates that this PR will close issue #123.
-->

N/A

## Rationale for this change

<!--
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.
-->

When hashing list arrays we hash all bytes of the child array, even if
we slice to a certain range of values. Refactor to slice only the needed
bytes; also do some general refactors.

## 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.
-->

Refactor list array hashing.

## Are these changes tested?

<!--
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)?
-->

Added test.

## Are there any user-facing changes?

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

No.

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

Author: Jefffrey

Cargo.lock

@@ -72,6 +72,7 @@ half = { workspace = true }
 hashbrown = { workspace = true }
 hex = { workspace = true, optional = true }
 indexmap = { workspace = true }
+itertools = { workspace = true }
 libc = "0.2.180"
 log = { workspace = true }
 object_store = { workspace = true, optional = true }

datafusion/common/Cargo.toml

@@ -23,6 +23,7 @@ use arrow::array::*;
 use arrow::datatypes::*;
 #[cfg(not(feature = "force_hash_collisions"))]
 use arrow::{downcast_dictionary_array, downcast_primitive_array};
+use itertools::Itertools;
 
 #[cfg(not(feature = "force_hash_collisions"))]
 use crate::cast::{
@@ -514,24 +515,41 @@ fn hash_list_array<OffsetSize>(
 where
     OffsetSize: OffsetSizeTrait,
 {
-    let values = array.values();
-    let offsets = array.value_offsets();
-    let nulls = array.nulls();
-    let mut values_hashes = vec![0u64; values.len()];
-    create_hashes([values], random_state, &mut values_hashes)?;
-    if let Some(nulls) = nulls {
-        for (i, (start, stop)) in offsets.iter().zip(offsets.iter().skip(1)).enumerate() {
-            if nulls.is_valid(i) {
+    // In case values is sliced, hash only the bytes used by the offsets of this ListArray
+    let first_offset = array.value_offsets().first().cloned().unwrap_or_default();
+    let last_offset = array.value_offsets().last().cloned().unwrap_or_default();
+    let value_bytes_len = (last_offset - first_offset).as_usize();
+    let mut values_hashes = vec![0u64; value_bytes_len];
+    create_hashes(
+        [array
+            .values()
+            .slice(first_offset.as_usize(), value_bytes_len)],
+        random_state,
+        &mut values_hashes,
+    )?;
+
+    if array.null_count() > 0 {
+        for (i, (start, stop)) in array.value_offsets().iter().tuple_windows().enumerate()
+        {
+            if array.is_valid(i) {
                 let hash = &mut hashes_buffer[i];
-                for values_hash in &values_hashes[start.as_usize()..stop.as_usize()] {
+                for values_hash in &values_hashes[(*start - first_offset).as_usize()
+                    ..(*stop - first_offset).as_usize()]
+                {
                     *hash = combine_hashes(*hash, *values_hash);
                 }
             }
         }
     } else {
-        for (i, (start, stop)) in offsets.iter().zip(offsets.iter().skip(1)).enumerate() {
-            let hash = &mut hashes_buffer[i];
-            for values_hash in &values_hashes[start.as_usize()..stop.as_usize()] {
+        for ((start, stop), hash) in array
+            .value_offsets()
+            .iter()
+            .tuple_windows()
+            .zip(hashes_buffer.iter_mut())
+        {
+            for values_hash in &values_hashes
+                [(*start - first_offset).as_usize()..(*stop - first_offset).as_usize()]
+            {
                 *hash = combine_hashes(*hash, *values_hash);
             }
         }
@@ -1176,6 +1194,30 @@ mod tests {
         assert_eq!(hashes[1], hashes[6]); // null vs empty list
     }
 
+    #[test]
+    #[cfg(not(feature = "force_hash_collisions"))]
+    fn create_hashes_for_sliced_list_arrays() {
+        let data = vec![
+            Some(vec![Some(0), Some(1), Some(2)]),
+            None,
+            // Slice from here
+            Some(vec![Some(3), None, Some(5)]),
+            Some(vec![Some(3), None, Some(5)]),
+            None,
+            // To here
+            Some(vec![Some(0), Some(1), Some(2)]),
+            Some(vec![]),
+        ];
+        let list_array =
+            Arc::new(ListArray::from_iter_primitive::<Int32Type, _, _>(data)) as ArrayRef;
+        let list_array = list_array.slice(2, 3);
+        let random_state = RandomState::with_seeds(0, 0, 0, 0);
+        let mut hashes = vec![0; list_array.len()];
+        create_hashes(&[list_array], &random_state, &mut hashes).unwrap();
+        assert_eq!(hashes[0], hashes[1]);
+        assert_ne!(hashes[1], hashes[2]);
+    }
+
     #[test]
     // Tests actual values of hashes, which are different if forcing collisions
     #[cfg(not(feature = "force_hash_collisions"))]