Merge remote-tracking branch 'apache/main' into alamb/default_arc

Author: alamb

datafusion/common/src/hash_utils.rs

@@ -201,6 +201,7 @@ hash_float_value!((half::f16, u16), (f32, u32), (f64, u64));
 /// Create a `SeedableRandomState` whose per-hasher seed incorporates `seed`.
 /// This folds the previous hash into the hasher's initial state so only the
 /// new value needs to pass through the hash function — same cost as `hash_one`.
+#[cfg(not(feature = "force_hash_collisions"))]
 #[inline]
 #[cfg(not(feature = "force_hash_collisions"))]
 fn seeded_state(seed: u64) -> foldhash::fast::SeedableRandomState {
@@ -307,6 +308,7 @@ fn hash_array<T>(
 /// HAS_NULLS: do we have to check null in the inner loop
 /// HAS_BUFFERS: if true, array has external buffers; if false, all strings are inlined/ less then 12 bytes
 /// REHASH: if true, combining with existing hash, otherwise initializing
+#[cfg(not(feature = "force_hash_collisions"))]
 #[inline(never)]
 #[cfg(not(feature = "force_hash_collisions"))]
 fn hash_string_view_array_inner<
@@ -434,6 +436,7 @@ fn hash_generic_byte_view_array<T: ByteViewType>(
 /// - `HAS_NULL_KEYS`: Whether to check for null dictionary keys
 /// - `HAS_NULL_VALUES`: Whether to check for null dictionary values
 /// - `MULTI_COL`: Whether to combine with existing hash (true) or initialize (false)
+#[cfg(not(feature = "force_hash_collisions"))]
 #[inline(never)]
 #[cfg(not(feature = "force_hash_collisions"))]
 fn hash_dictionary_inner<

datafusion/common/src/pruning.rs

@@ -121,6 +121,7 @@ pub trait PruningStatistics {
     /// container, return `None` (the default).
     ///
     /// Note: the returned array must contain [`Self::num_containers`] rows
+    #[allow(clippy::allow_attributes, clippy::mutable_key_type)] // ScalarValue has interior mutability but is intentionally used as hash key
     fn contained(
         &self,
         column: &Column,
@@ -526,6 +527,7 @@ impl PruningStatistics for CompositePruningStatistics {
 
 #[cfg(test)]
 #[expect(deprecated)]
+#[allow(clippy::allow_attributes, clippy::mutable_key_type)] // ScalarValue has interior mutability but is intentionally used as hash key
 mod tests {
     use crate::{
         ColumnStatistics,

datafusion/common/src/scalar/mod.rs

@@ -4601,6 +4601,7 @@ impl ScalarValue {
     /// Estimates [size](Self::size) of [`HashSet`] in bytes.
     ///
     /// Includes the size of the [`HashSet`] container itself.
+    #[allow(clippy::allow_attributes, clippy::mutable_key_type)] // ScalarValue has interior mutability but is intentionally used as hash key
     pub fn size_of_hashset<S>(set: &HashSet<Self, S>) -> usize {
         size_of_val(set)
             + (size_of::<ScalarValue>() * set.capacity())
@@ -7263,6 +7264,8 @@ mod tests {
             size_of::<Vec<ScalarValue>>() + (9 * size_of::<ScalarValue>()) + sv_size,
         );
 
+        #[allow(clippy::allow_attributes, clippy::mutable_key_type)]
+        // ScalarValue has interior mutability but is intentionally used as hash key
         let mut s = HashSet::with_capacity(0);
         // do NOT clone `sv` here because this may shrink the vector capacity
         s.insert(v.pop().unwrap());

datafusion/core/src/physical_planner.rs

@@ -2093,6 +2093,7 @@ fn get_physical_expr_pair(
 /// A vector of unqualified filter expressions that can be passed to the TableProvider for execution.
 /// Returns an empty vector if no applicable filters are found.
 ///
+#[allow(clippy::allow_attributes, clippy::mutable_key_type)] // Expr contains Arc with interior mutability but is intentionally used as hash key
 fn extract_dml_filters(
     input: &Arc<LogicalPlan>,
     target: &TableReference,

datafusion/datasource/src/file_groups.rs

@@ -488,6 +488,7 @@ impl FileGroup {
     ///
     /// Note: May return fewer groups than `max_target_partitions` when the
     /// number of unique partition values is less than the target.
+    #[allow(clippy::allow_attributes, clippy::mutable_key_type)] // ScalarValue has interior mutability but is intentionally used as hash key
     pub fn group_by_partition_values(
         self,
         max_target_partitions: usize,

datafusion/expr/src/logical_plan/builder.rs

@@ -2132,6 +2132,8 @@ pub fn wrap_projection_for_join_if_necessary(
             .into_iter()
             .map(Expr::Column)
             .collect::<Vec<_>>();
+        #[allow(clippy::allow_attributes, clippy::mutable_key_type)]
+        // Expr contains Arc with interior mutability but is intentionally used as hash key
         let join_key_items = alias_join_keys
             .iter()
             .flat_map(|expr| expr.try_as_col().is_none().then_some(expr))

datafusion/functions-aggregate/src/median.rs

@@ -285,6 +285,7 @@ impl<T: ArrowNumericType> Accumulator for MedianAccumulator<T> {
         size_of_val(self) + self.all_values.capacity() * size_of::<T::Native>()
     }
 
+    #[allow(clippy::allow_attributes, clippy::mutable_key_type)] // ScalarValue has interior mutability but is intentionally used as hash key
     fn retract_batch(&mut self, values: &[ArrayRef]) -> Result<()> {
         let mut to_remove: HashMap<ScalarValue, usize> = HashMap::new();
 

datafusion/functions-aggregate/src/percentile_cont.rs

@@ -440,6 +440,7 @@ where
         size_of_val(self) + self.all_values.capacity() * size_of::<T::Native>()
     }
 
+    #[allow(clippy::allow_attributes, clippy::mutable_key_type)] // ScalarValue has interior mutability but is intentionally used as hash key
     fn retract_batch(&mut self, values: &[ArrayRef]) -> Result<()> {
         let mut to_remove: HashMap<ScalarValue, usize> = HashMap::new();
         for i in 0..values[0].len() {

datafusion/optimizer/src/analyzer/resolve_grouping_function.rs

@@ -72,6 +72,7 @@ fn group_expr_to_bitmap_index(group_expr: &[Expr]) -> Result<HashMap<&Expr, usiz
         .collect::<HashMap<_, _>>())
 }
 
+#[allow(clippy::allow_attributes, clippy::mutable_key_type)] // Expr contains Arc with interior mutability but is intentionally used as hash key
 fn replace_grouping_exprs(
     input: Arc<LogicalPlan>,
     schema: &DFSchema,
@@ -158,6 +159,7 @@ fn contains_grouping_function(exprs: &[Expr]) -> bool {
 }
 
 /// Validate that the arguments to the grouping function are in the group by clause.
+#[allow(clippy::allow_attributes, clippy::mutable_key_type)] // Expr contains Arc with interior mutability but is intentionally used as hash key
 fn validate_args(
     function: &AggregateFunction,
     group_by_expr: &HashMap<&Expr, usize>,
@@ -178,6 +180,7 @@ fn validate_args(
     }
 }
 
+#[allow(clippy::allow_attributes, clippy::mutable_key_type)] // Expr contains Arc with interior mutability but is intentionally used as hash key
 fn grouping_function_on_id(
     function: &AggregateFunction,
     group_by_expr: &HashMap<&Expr, usize>,

datafusion/optimizer/src/scalar_subquery_to_join.rs

@@ -144,7 +144,11 @@ impl OptimizerRule for ScalarSubqueryToJoin {
                 }
 
                 let mut all_subqueries = vec![];
+                #[allow(clippy::allow_attributes, clippy::mutable_key_type)]
+                // Expr contains Arc with interior mutability but is intentionally used as hash key
                 let mut expr_to_rewrite_expr_map = HashMap::new();
+                #[allow(clippy::allow_attributes, clippy::mutable_key_type)]
+                // Expr contains Arc with interior mutability but is intentionally used as hash key
                 let mut subquery_to_expr_map = HashMap::new();
                 for expr in projection.expr.iter() {
                     let (subqueries, rewrite_exprs) =