Bring dynamic filtering cherry picks (#70)

* Fix dynamic filter is_used function (apache#19734)

## 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 apache#123` indicates that this PR will close issue apache#123.
-->

- Closes apache#19715.

## Rationale for this change

The:is_used() API incorrectly returned false for custom `DataSource`
implementations that didn't call reassign_expr_columns() ->
with_new_children() . This caused `HashJoinExec` to skip computing
dynamic filters even when they were actually being used.

## What changes are included in this PR?

Updated is_used() to check both outer and inner Arc counts

## Are these changes tested?

Functionality is covered by existing test
`test_hashjoin_dynamic_filter_pushdown_is_used`. I was not sure if to
add a repro since it would require adding a custom `DataSource`, the
current tests in
datafusion/core/tests/physical_optimizer/filter_pushdown/mod.rs use
`FileScanConfig`

## Are there any user-facing changes?

no

(cherry picked from commit 278950a)

* Simplify wait_complete function (apache#19937)

## Which issue does this PR close?

## Rationale for this change

The current v52 signature `pub async fn wait_complete(self: &Arc<Self>)`
(introduced in apache#19546) is a bit unergonomic. The method requires
`&Arc<DynamicFilterPhysicalExpr>`, but when working with `Arc<dyn
PhysicalExpr>`, downcasting only gives you `&DynamicFilterPhysicalExpr`.
Since you can't convert `&DynamicFilterPhysicalExpr` to
`Arc<DynamicFilterPhysicalExpr>`, the method becomes impossible to call.

The `&Arc<Self>` param was used to check` is_used()` via Arc strong
count, but this was overly defensive.

## What changes are included in this PR?

- Changed `DynamicFilterPhysicalExpr::wait_complete` signature from `pub
async fn wait_complete(self: &Arc<Self>)` to `pub async fn
wait_complete(&self)`.

- Removed the `is_used()` check from `wait_complete()` - this method,
like `wait_update()`, should only be called on filters that have
consumers. If the caller doesn't know whether the filter has consumers,
they should call `is_used()` first to avoid waiting indefinitely. This
approach avoids complex signatures and dependencies between the APIs
methods.

## Are these changes tested?

Yes, existing tests cover this functionality, I removed the "mock"
consumer from `test_hash_join_marks_filter_complete_empty_build_side`
and `test_hash_join_marks_filter_complete` since the fix in
apache#19734 makes is_used check the
outer struct `strong_count` as well.

## Are there any user-facing changes?

The signature of `wait_complete` changed.

(cherry picked from commit bef1368)

Author: LiaCastaneda

datafusion/physical-expr/src/expressions/dynamic_filters.rs

@@ -272,6 +272,10 @@ impl DynamicFilterPhysicalExpr {
     ///
     /// This method will return when [`Self::update`] is called and the generation increases.
     /// It does not guarantee that the filter is complete.
+    ///
+    /// Producers (e.g.) HashJoinExec may never update the expression or mark it as completed if there are no consumers.
+    /// If you call this method on a dynamic filter created by such a producer and there are no consumers registered this method would wait indefinitely.
+    /// This should not happen under normal operation and would indicate a programming error either in your producer or in DataFusion if the producer is a built in node.
     pub async fn wait_update(&self) {
         let mut rx = self.state_watch.subscribe();
         // Get the current generation
@@ -283,17 +287,16 @@ impl DynamicFilterPhysicalExpr {
 
     /// Wait asynchronously until this dynamic filter is marked as complete.
     ///
-    /// This method returns immediately if the filter is already complete or if the filter
-    /// is not being used by any consumers.
+    /// This method returns immediately if the filter is already complete.
     /// Otherwise, it waits until [`Self::mark_complete`] is called.
     ///
     /// Unlike [`Self::wait_update`], this method guarantees that when it returns,
     /// the filter is fully complete with no more updates expected.
-    pub async fn wait_complete(self: &Arc<Self>) {
-        if !self.is_used() {
-            return;
-        }
-
+    ///
+    /// Producers (e.g.) HashJoinExec may never update the expression or mark it as completed if there are no consumers.
+    /// If you call this method on a dynamic filter created by such a producer and there are no consumers registered this method would wait indefinitely.
+    /// This should not happen under normal operation and would indicate a programming error either in your producer or in DataFusion if the producer is a built in node.
+    pub async fn wait_complete(&self) {
         if self.inner.read().is_complete {
             return;
         }
@@ -310,14 +313,14 @@ impl DynamicFilterPhysicalExpr {
     /// that created the filter). This is useful to avoid computing expensive filter
     /// expressions when no consumer will actually use them.
     ///
-    /// Note: We check the inner Arc's strong_count, not the outer Arc's count, because
-    /// when filters are transformed (e.g., via reassign_expr_columns during filter pushdown),
-    /// new outer Arc instances are created via with_new_children(), but they all share the
-    /// same inner `Arc<RwLock<Inner>>`. This is what allows filter updates to propagate to
-    /// consumers even after transformation.
+    /// # Implementation Details
+    ///
+    /// We check both Arc counts to handle two cases:
+    /// - Transformed filters (via `with_new_children`) share the inner Arc (inner count > 1)
+    /// - Direct clones (via `Arc::clone`) increment the outer count (outer count > 1)
     pub fn is_used(self: &Arc<Self>) -> bool {
         // Strong count > 1 means at least one consumer is holding a reference beyond the producer.
-        Arc::strong_count(&self.inner) > 1
+        Arc::strong_count(self) > 1 || Arc::strong_count(&self.inner) > 1
     }
 
     fn render(

datafusion/physical-plan/src/joins/hash_join/exec.rs

@@ -513,10 +513,8 @@ impl HashJoinExec {
     ///
     /// This method is intended for testing only and should not be used in production code.
     #[doc(hidden)]
-    pub fn dynamic_filter_for_test(&self) -> Option<Arc<DynamicFilterPhysicalExpr>> {
-        self.dynamic_filter
-            .as_ref()
-            .map(|df| Arc::clone(&df.filter))
+    pub fn dynamic_filter_for_test(&self) -> Option<&Arc<DynamicFilterPhysicalExpr>> {
+        self.dynamic_filter.as_ref().map(|df| &df.filter)
     }
 
     /// Calculate order preservation flags for this hash join.
@@ -4635,11 +4633,6 @@ mod tests {
         let dynamic_filter = HashJoinExec::create_dynamic_filter(&on);
         let dynamic_filter_clone = Arc::clone(&dynamic_filter);
 
-        // Simulate a consumer by creating a transformed copy (what happens during filter pushdown)
-        let _consumer = Arc::clone(&dynamic_filter)
-            .with_new_children(vec![])
-            .unwrap();
-
         // Create HashJoinExec with the dynamic filter
         let mut join = HashJoinExec::try_new(
             left,
@@ -4688,11 +4681,6 @@ mod tests {
         let dynamic_filter = HashJoinExec::create_dynamic_filter(&on);
         let dynamic_filter_clone = Arc::clone(&dynamic_filter);
 
-        // Simulate a consumer by creating a transformed copy (what happens during filter pushdown)
-        let _consumer = Arc::clone(&dynamic_filter)
-            .with_new_children(vec![])
-            .unwrap();
-
         // Create HashJoinExec with the dynamic filter
         let mut join = HashJoinExec::try_new(
             left,